Ownership Structure and Executive Compensation in Canadian ... · Ownership Structure and Executive...

Ownership Structure and Executive Compensation in Canadian Corporations

A Thesis Submitted to the College of

Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the Degree of Master of Science in Finance

In the Department of Finance and Management Science

Edwards School of Business

University of Saskatchewan

Saskatoon, Saskatchewan, Canada

By

WEIWEI JIANG

Copyright Weiwei Jiang, April, 2011. All rights reserved.

i

Permission to Use

In presenting this thesis in partial fulfilment of the requirements for a Postgraduate

degree from the University of Saskatchewan, I agree that the Libraries of this University may

make it freely available for inspection. I further agree that permission for copying of this thesis

in any manner, in whole or in part, for scholarly purposes may be granted by the professor or

professors who supervised my thesis work or, in their absence, by the Head of the Department or

the Dean of the College in which my thesis work was done. It is understood that any copying or

publication or use of this thesis or parts thereof for financial gain shall not be allowed without

my written permission. It is also understood that due recognition shall be given to me and to the

University of Saskatchewan in any scholarly use which may be made of any material in my

thesis.

Requests for permission to copy or to make other use of material in this thesis in whole or

part should be addressed to:

Head of the Department of Finance and Management Science

Edwards School of Business

University of Saskatchewan

25 Campus Drive

Saskatoon, Saskatchewan S7N 5A7

ii

Abstract

Agency theory, proposed by previous studies such as Guidry, Leone, and Rock (1999)

and Arya and Huey-Lian (2004), suggests that bonus and other accounting-metric-based

compensation can motivate managers to perform well in the short horizon while equity-based

compensation, such as restricted shares and stock options, can serve the purpose of aligning the

long run interests of shareholders and managers. The empirical evidence, for example Jensen and

Murphy (1990), Kaplan (1994), Hall and Liebman (1998), Murphy (1999), Zhou (2000), and

Chowdhury and Wang (2009), confirms that incentive compensation is popular in many

countries. However, recent studies suggest that the relation between performance and incentive

compensation is weak. Shaw and Zhang (2010) find that CEO bonus compensation is less

sensitive to poor earnings performance than it is to good earnings performance. Fahlenbrach and

Stulz (2011) study the relation between bank performance during the 2008 bank crisis and the

bonus and equity-based compensation of bank CEOs. They find that banks with CEOs whose

incentives were better aligned with the interests of shareholders performed worse than other

banks.

This study examines whether ownership structure can explain the differences among

compensation structures of chief executive officers (CEOs). In particular, we examine the

compensation structure of three distinct groups: family-controlled, institution-controlled, and

widely-held firms. We distinguish these three kinds of firms to represent different levels of

market imperfection. Compared with family-controlled and institution-controlled firms, widely

held firms have dispersed ownership. The most significant weakness of a widely-held ownership

structure is the lack of shareholder monitoring due to the unmatched benefit and cost of

monitoring for small shareholders. In contrast, a holder of a large block of shares will have the

same monitoring costs but the benefits to this shareholder from monitoring management and

reducing agency costs would be substantial and larger than the costs of monitoring. Thus the

presence of a large shareholder will reduce the agency costs. In addition, large shareholders may

be willing to spend time and effort continuously to collect more information on management

performance or to estimate the firm’s investment projects. This behaviour will reduce the

problems that arise from information asymmetry and will decrease the waste of free cash flows

by managers.

iii

Both family-controlled firms and institution-controlled firms have large shareholders.

However, whether or not the control shareholders are playing an active monitoring role is still an

important issue. From the viewpoint of aligning the interests of managers and shareholders, the

family-controlled group is superior to the institution-controlled group. First, institutions are more

flexible in moving their ownership from one firm to another depending on performance. If the

costs of monitoring are high in comparison to the costs of rebalancing portfolios, institutions will

choose to rebalance instead of monitoring. In contrast, a family that controls a firm does not have

this flexibility. Second, family-controlled firms generally assign influential positions to family

members whose focus is in line with that of the family group. Even though a non family member

may be appointed as the manager, the level of monitoring is significant given the high ownership

concentration by the family. However, the level of monitoring by a family may not necessarily

translate into a reduction of agency costs for minority shareholders. Indeed, previous studies

suggest that significant family ownership may lead to agency costs of its own. The family may

divert company resources for its own benefit despite the presence of a manager who may or may

not be a family member. Essentially, the family and the manager can collude to spend on perks

and personal benefits at the expense of minority shareholders. Chourou (2010) suggests that

excessive compensation of chief executive officers at some family owned Canadian corporations

may be viewed as expropriation of minority rights.

Overall, the main objective of this study is to examine whether block-holder monitoring

is a substitute to the incentive components of compensation. We propose that as we move from

widely-held to institution-controlled the level of monitoring may or may not increase. However,

as we move further into higher control, as may be suggested by family ownership, the level of

monitoring will increase but this monitoring may not necessarily reduce agency costs. The

results show that the institution-controlled firms pay significantly less bonus compensation per

dollar of assets than widely-held firms but the differences in equity based compensation are not

significant. In addition, the family-controlled corporations offer the lowest performance-based

compensation, bonus per dollar of assets, in comparison to the institution-controlled and the

widely-held groups. These results indicate that the family-controlled Canadian corporations rely

more on monitoring managers than paying them incentive payments in the form of bonus

payments. In addition, our results indicate that the institutions which control corporations may be

monitoring the managers of these corporations but this monitoring does not significantly reduce

iv

the need for the long-term incentive components of compensation. This result suggests that

institutions may monitor the short-term performance effectively but they may prefer rebalancing

their portfolio rather than monitoring long term performance.

v

ACKNOWLEDGMENTS

My profound appreciation and gratitude goes to my supervisors Professor George

Tannous and Dr. Fan Yang first for being an admirable coach in guiding me through my thesis. I

sincerely acknowledge their rigorous commitment and endless effort to ensure the quality of this

thesis. They in depth knowledge, intellectual ability and generosity have helped me to have a

good understanding of the Corporate Governance issues. Without the large amount of time and

energy on helping me from them, this thesis could never have been completed. It is my pleasure,

honor, and luck to have such a great opportunity to learn from them. I also would like to dedicate

sincere appreciation to my supervisory committee member Dr. Zhuyu Wu, as well as my external

examiner, Dr. Don Cyr, for their insight comments and suggestions.

I am thankful to all my professors to their contributions to my knowledge in Finance. I

am thankful to the countless help and support that I received from Dr. Marie Racine. I am

thankful to Dr. Abdullah Mamun, Dr. Craig Wilson, and Dr. Dev Mishra for their contributions

to my knowledge in Finance. I am thankful to Ms. Brenda Orischuk, for her assistance during my

studies at the University of Saskatchewan. I acknowledge the encouragements and help of my

friends and classmates. Special thanks to Yuting Fu, Eva Yang, Mo Zhou, and Lucy Li.

Finally, my deepest love and gratitude go to my best parents and my husband for their

unconditional love, encourage, and understanding.

vi

TABLE OF CONTENTS

TABLE OF CONTENTS ..................................................................................................... i ABSTRACT ........................................................................................................................ ii ACKONWLEDGEMENTS ................................................................................................. v LIST OF FIGURES ......................................................................................................... viii LIST OF TABLES .............................................................................................................. ix CHAPTER 1 INTRODUCTION ............................................................................................................... 1 CHAPTER 2 LITERATURE REVIEW ....................................................................................................6

2.1 The Components of CEO Compensation .................................................................. 6 2.2 Agency Theory and Incentive Compensation ........................................................... 7 2.3 Empirical Evidence Regarding the Relation between Pay and Performance ........... 8 2.4 Ownership Structure .............................................................................................. .10

2.4.1 Widely-held Group Versus the Concentrated Group………………………....12 2.4.1 Family-controlled Group Versus Institution-controlled Group……………....14

2.5 Other Factors ........................................................................................................... 17

CHAPTER 3 THEORETICAL ARGUMENTS AND HYPOTHESES ..................................................19 CHAPTER 4 DATA ................................................................................................................................23

4.1 Ownership Structure ............................................................................................... 23 4.2 CEO Compensation ................................................................................................ 24 4.3 Variables ................................................................................................................ 27

4.3.1 Dependent Variables……………………………………………………….....27 4.3.2 Control Variables……………………………………………..……………... 27 4.3.3 DummyVariables……………………………………………..……………... 28

CHAPTER 5 DESCRIPTIVE STATISTICS AND UNIVARIATE TESTS ...........................................30

5.1 Permanment CEO Firms ......................................................................................... 30 5.2 Transient CEO Group ............................................................................................. 32 5.3 Aggregate Sample ................................................................................................... 35

5.3.1 Descriptive Statistics....……………………………………………………....35 5.3.2 Univarite Variables……………………………………………..…………....36

CHAPTER 6 MULTIVARIATE ANALYSIS.........................................................................................39

6.1 Natural Logarithm of Compensation as Dependent Variable ................................. 41 6.2 Bonus and Incentive Compensation as Percentage of Total Assets ....................... 43

6.2.1 Annual Bonus as a Percentage of Total Assets…………………………….....44

vii

6.2.2 Contingent Compensation as a Percentage of Total Assets…………….….....45 6.3 Bonus and Incentive Compensation as Percentage of Total Assets ....................... 46

6.3.1 Annual Bonus as a Percentage of Total Pay……………………………….....46 6.3.2 Contingent Compensation as a Percentage of Total Pay…………………......47

6.4 The Relation between Compensation and Total Market Return ............................. 48

CHAPTER 7 CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER RESEARCH ......... 51

REFERENCES ................................................................................................................. 55

viii

LIST OF FIGURES

Figure 1.1: Sarlay in monetary terms paid by permanent CEO firms .................................90

Figure 1.2: Bonus in monetary terms paid by permanent CEO firms ..................................90

Figure 1.3: Contingent compensation in monetary terms paid by permanent CEO firms ....91

Figure 1.3: Total compensation in monetary terms paid by permanent CEO firms .............91

Figure 2.1: Salary as a percentage of total compensation paid by permanent CEO firms ...92

Figure 2.2: Annual Bonus as a percentage of total compensation paid by permanent CEO firms..............................................................................................................................92

Figure 2.3: Contingent Compensation as a percentage of total compensation paid by permanent CEO firms ............................................................................................................93

Figure 3.1: Changes in compensation following CEO turnovers in family-controlled firms..............................................................................................................................94

Figure 3.2: Changes in compensation following CEO turnovers in institution-controlled firms..............................................................................................................................94

Figure 3.3: Changes in compensation following CEO turnovers in widely-held firms ........95

Figure 4.1: Changes in compensation following CEO retirements in family-controlled firms..............................................................................................................................96

Figure 4.2: Changes in compensation following CEO retirements in institution-controlled firms..............................................................................................................................96

Figure 5.1: Total compensation as a function of total assets (all data) .................................97

Figure 5.2: Total compensation as a function of total assets (assets sizes of $20.48 million-9 billion) ..................................................................................................................98

Figure 5.3: Total compensation as a function of total assets (assets sizes of $9.1 billion-$55 billion) ..................................................................................................................99

Figure 5.4: Total compensation as a function of total assets (asset sizes larger than $55 billion)..............................................................................................................................100

Figure 6.1: Salary and Bonus are as a function of total assets (all data) ..............................101

Figure 6.2: Salary and Bonus are as a function of total assets (assets sizes of $20.48 million-55 billion) ..................................................................................................................102

ix

LIST OF TABLES

Table 1.1: Mean value of CEO monetary compensation in permanent CEOs firms .............61

Table 1.2: Mean value of salary, annual bonus, and contingent pay as a percentage of total pay in permanent CEOs firms .........................................................................................61

Table 2.1: Descriptive statistics in the transient CEO group .................................................62

Table 2.2: Comparison of the compensation of incoming CEOs and that of their predecessors..............................................................................................................................63

Table 3: Descriptive statistics related to the dependent and control variables ......................64

Table 4: Descriptive statistics related to the dependent and control variables in three different ownership structures ............................................................................................65

Table 5: Descriptive statistics related to the dependent and control variables in three different industries ..............................................................................................................66

Table 6.1: T-test: two sample assuming unequal variances ...................................................67

Table 6.2: Descriptive statistics mean value annual bonus and contingent pay ....................67

Table 7: Descriptive statistics mean for the control variables ...............................................68

Table 8.1: Correlation of variables in the aggressive sample ................................................69

Table 8.2: Correlation of variables in the family-controlled firms ........................................70

Table 8.3: Correlation of variables in the institution-controlled firms ..................................71

Table 8.4: Correlation of variables in the widely-held firms .................................................72

Table 9.1: The impact of ownership structure on the natural log of annual bonus (OLS) ....73

Table 9.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) .........................................................................................74

Table 10.1: The impact of ownership structure on the natural log of contingent compensation (OLS) ...................................................................................................................75

Table 10.2: Natural log of contingent compensation in widely-held, institution-controlled, and family-controlled firms (OLS) .............................................................................76

Table 11.1: The impact of ownership structure on the annual bonus as a proportion of total assets (Tobit) .................................................................................................................77

x

Table 11.2: Annual bonus as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) .....................................................................78

Table 12.1: The impact of ownership structure on the contingent compensation as a proportion of total assets (Tobit) ................................................................................................79

Table 12.2: Contingent compensation as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) ...................................................80

Table 13.1: The impact of ownership structure on the annual bonus as a proportion of total pay (Tobit) ..................................................................................................................81

Table 13.2: Annual bonus as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) ............................................................................82

Table 14.1: The impact of ownership structure on the contingent compensation as a proportion of total pay (Tobit) ...................................................................................................83

Table 14.2: Contingent pay as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) ............................................................................84

Table 15: Correlation of variables in the family-controlled, institution-controlled, and widely-held firms .............................................................................................................85

Table 16.1: The impact of ownership structure on the natural log of annual bonus (OLS) ..86

Table 16.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) .........................................................................................87

Table 17.1: The impact of ownership structure on the natural log of contingent pay (OLS) 88

Table 17.2: Natural log of contingent compensation in widely-held, institution-controlled, and family-controlled firms (OLS) .............................................................................89

1

CHAPTER 1 Introduction

In theory, efficient pay contracts bond executive compensation with firm performance,

and offer strong incentives for executives to act in shareholders’ best interests. Guidry, Leone,

and Rock (1999) suggest that bonus and other accounting-metric-based compensation can

motivate managers to perform well in the short run. Jenson and Murphy (1990) propose that

CEO ownership of their firm’s stock is the largest CEO performance incentives. Lamber,

Larchker, and Verrecchia (1991) point out that the stock-based compensation can mitigate

agency problems. Similarly, Arya and Huey-Lian (2004) propose that equity-based

compensation, such as restricted shares and stock options, can align the long-term interests of

shareholders and managers.

A number of empirical studies confirm that incentive pay, including both short-term

bonus and long-term equity-based compensation, are used to reduce agency costs. Kaplan (1994)

provides evidence suggesting that the fortunes of Japanese top executives are related to stock

performance and to factors that are conducive to stock and earning performance. Hall and

Liebman (1998) provide evidence suggesting that firm performance is correlated to CEO

compensation. Murphy (1999) reports that the relation between compensation and performance

in the United States is stronger than the same relation in other countries. In the context of Canada,

Zhou (2000) suggests that executive compensation is positively correlated to firm performance

with an overall weak relationship. Another Canadian study, Chowdhury and Wang (2009), find

that contingent pay and its ratio to total pay have been increasing in Canada from 1995 to 2002.

Recently, some studies show that the relation between performance and incentive

compensation is weak. Shaw and Zhang (2010) find that CEO bonus compensation is less

sensitive to poor earnings performance than it is to good earnings performance. They suggest

that CEOs get rewards even with poor firm performance. Similarly, Fahlenbrach and Stulz (2011)

find no evidence to support the proposition that banks with CEOs whose incentives were not

well aligned with the interests of their shareholders performed worse.

This study examines whether ownership structure can explain the differences among the

levels and structures of chief executive officer (CEO) compensation. In particular, we examine

the compensation of three distinct groups: family-controlled, institution-controlled, and widely-

held firms. We distinguish these three kinds of firms to represent different levels control and

2

monitoring provided by widely disbursed and concentrated ownership and by institutions as

opposed to families.

Previous studies suggest that the most significant weakness of a widely-held ownership

structure is the lack of shareholder monitoring due to the unmatched benefit and cost of

monitoring for small shareholders. Demsetz (1983) suggests that when ownership is widely

dispersed across many individuals and institutions, shareholders cannot exercise real power to

oversee managerial performance in modern corporations. The existence of at least one large

shareholder will reduce the agency costs and asymmetric information. McConaughy et al. (1998)

examine the efficiency, measured by sales growth, and value of family-controlled firms. Their

family-controlled firms are defined as public corporations whose CEOs are either the founder or

a member of the founder’s family. Controlling for size, industry and ownership effects, they

apply a matched-pairs methodology. Their key finding is that family controlled firms are more

valuable and efficient than firms of the same size, in the same industry, and with similar

managerial ownership. Their findings also emphasize that who owns the shares is more

important than ownership concentration. Firth, Fung and Rui (2006) indicate that concentrated

ownership reduces agency costs.


David, Kochhar, and Levitas (1998) argue that institutions have the obligation to know and

protect what they invest. They should take proactive actions, so that the managements of investee

firms work towards maximizing shareholder value. We propose that firms with a concentrated

ownership structure would behave differently depending on whether they are institution-

controlled or family-controlled. Shleifer and Vishny (1986) show that institution block holders

do not usually interfere with management but they perform better monitoring than small

shareholders in widely-held firms. However, whether or not institutions are effective in

controlling managers is still an important issue. Interviews with six investment managers that

control significant pension assets in Canada reveal that institutions actively communicate with

and monitor managements of the firms they invest in but these investment managers stopped

short of claiming that they attempt to exercise control over managements. Institutions are more

flexible in moving their ownership from one firm to another depending on performance. If the

costs of monitoring are high in comparison to the costs of rebalancing portfolios, institutions will

choose to rebalance instead of trying to change managerial attitudes and decisions. Therefore, we

3

propose that institutions are better positioned than individuals to monitor managers and provide

recommendations but they are not likely to exert significant and effective control over the

managers of the firms in which they have control.

In contrast, a family that holds a controlling portion of voting shares is likely to have a

significant and personal interest in the firm. They family would generally assign influential

positions to family members or to managers who are controlled indirectly by the family. Thus,

the managers of family-controlled firms are likely to be monitored more closely than the

managers of institution-controlled firms and they are more likely to be directed and influenced

by the controlling family. Many studies suggest that the fortunes of the managers of family-

controlled firms are very much tied with the fortunes of the families that control their firms.

Fama and Jensen (1983) and Demsetz (1983) suggest that a manager who cooperates with the

controlling family can guarantee employment at an attractive salary as long as the firm’s

performance is in line with the industry’s performance and the decisions of the manager are

consistent with the expectations of the family. Furthermore, Chen and Kensinger (1988) suggest

that managers of family-controlled firms may avoid risky ventures which might be desirable for

outside shareholders. Morck et al. (2000) suggest that the concentration of family wealth in a

business and the concern over the family legacy may explain why family-controlled firms may

display excessive risk-aversion and forego profitable expansion strategies and mergers. Chourou

(2010) suggests that excessive compensation of chief executive officers at some family owned

Canadian corporations may be a sign of cooperation between the controlling family and the CEO

and can be perceived as expropriation of minority rights. These arguments suggest that the

compensation packages of family-controlled managers are likely to be competitive in the market

for managerial talent, encourage good performance, promote cooperation with the controlling

family, and discourage managers from taking excessive risk. Therefore, we propose that the

incentive compensation in concentrated ownership firms may or may not vary depending on

whether the firm is family-controlled or institution-controlled. However, the relation between

incentive compensation and performance should be stronger when a family is the source of

ownership concentration.

Overall, we propose that incentive compensation may or may not vary across ownership

structures but the relation between incentive compensation and performance is likely to be weak

at the widely-held firms, stronger at institution-controlled firms, and strongest at the family-

4

controlled firms. Furthermore, we propose that if incentive compensation is different across the

ownership structures the differences are likely to be significant with the bonus component. We

suggest that the ownership structure that provides the strongest monitoring is likely to rely more

on bonus compensation as contingent component may reward managers due to factors beyond

their control. Overall, it is easier to see how managerial actions affect the measures upon which

bonuses are based rather than to point out the managerial actions that affect contingent

compensation.

We examine our theory regarding the relationship between ownership structure,

compensation, and monitoring by also considering how incoming CEOs are compensated in

comparison to their predecessors. Previous studies provide little information on the differences.

Ocasio (1994) suggests that CEO compensation is affected by the CEO’s tenure. In particular,

experience in the industry and in similar position may enable the incoming CEO to negotiate a

high compensation package and a structure that is in the best interests of the CEO. On the other

hand, the departure of a CEO could be seen as an opportunity for a firm to re-establish its own

priorities and to design the compensation package to promote the interests of shareholders and

the ultimate power brokers of the firm. Accordingly, we expect that the compensation of the

incoming CEOs would be structured differently than the compensation of their predecessors.

Furthermore, we propose that the structure of the compensation packages of incoming CEOs will

vary depending on the ownership structure.

Finally, we examine the relation between equity performance and compensation.

Previous studies find this relation to be positive but weak. We propose that separating firms

across ownership structures may reveal that the relation is significant for one ownership structure

and not significant for another. In particular, we propose that the relation is not likely to be

significant in the widely-held firms and institution-controlled firms but it is more likely to be

significant for firms in the family-controlled firms. In our view, the strength of the relation

between incentive compensation and equity performance should be inversely related to the level

of monitoring by the owners of the firms.

The remainder of this thesis is organized into seven sections. In Chapter 2, we review the

prior literature on ownership structures, CEO incentive compensation, and firm performance.

Theoretical arguments and hypothesis are discussed in Chapter 3. In Chapter 4, we describe the

data, define the variables, and explain the methods. In Chapter 5, we discuss our descriptive

5

statistics and results of univariate tests. In Chapter 6, we present and analyze the results of

multivariate tests. Chapter 7 covers conclusions and recommendations for future research.

6

CHPATER 2

Literature Review

This study examines whether ownership structure can explain the differences among

compensation structures of CEOs. In particular, we examine the compensation structure of three

distinct groups: family-controlled firms, institution-controlled firms and widely-held firms. Thus,

this chapter divides the literature into four sections: Section 2.1 discusses the components of

incentive pay. Section 2.2 reviews the literature that examines agency theory and incentive pay.

Section 2.3 provides a review of the empirical evidence regarding the relation between pay and

performance. Section 2.4 discusses the literature related to ownership structure and incentive pay,

and Section 2.5 presents some additional factors that may impact incentive pay.

2.1 The Components of CEO Compensation

Although structures of CEO compensation vary across firms and over time, previous

studies, for example Gray and Cannella (1997), find that CEO compensation contracts are

usually structured as a combination of cash compensation (salary and annual bonus) and long-

term compensation (stock-based and option-based compensation). Both the annual bonus and

long-term incentive pay are usually set based on some measures of firm performance. Yet, there

is a major difference between the two. The annual bonus is mostly based on accounting earnings

such as return on equity or return on assets while the long-term incentive pay is based on stock

returns.

Healy (1985) examines the nature of bonus payments. He notes that when actual firm

performance is below some minimum threshold, no funds are allocated to the bonus pool. As

firms perform better than the minimum, funds are linearly related to firm performance. When

firm performance is up to a ceiling, the bonus pool will be capped. He argues that since a large

part of CEO compensation is short-term bonuses based on accounting earnings, managers are

likely to choose to maximize their short-term bonuses.

Murphy (1999) reports that every profit-oriented company provides performance-based

bonus payments paid annually in cash. His findings suggest that if the CEO meets the

performance target, the CEO will receive the bonus which is usually determined as a given

percentage of his/her salary. Murphy (1999) finds that bonus contracts are usually written based

on accounting earnings and not explicitly on stock returns. Yet, he argues that annual bonus

7

payments have the potential of strengthening the alignment of interests between managers and

shareholders as they link the annual incentive awards to the future value of common shares.

Guidry, Leone, and Rock (1999) and Arya and Huey-Lian (2004), suggests that bonus and other

accounting-metric-based compensation can motivate the managers to perform well in the short

horizon while equity-based compensation, such as restricted shares and stock options, can serve

the purpose of aligning the long run interests of shareholders and managers.

2.2 Agency Theory and Incentive Compensation

Jensen and Meckling (1976) argue that an agency relationship arises when principals

appoint agents to make and execute decisions on behalf of the principals. Both the principal and

the agent want to maximize their respective utilities. Therefore, the agent will focus on

maximizing his/her own utility, not that of the principal. Unless the interests of the principal and

the agent are the same, the decisions of the agent will lead to suboptimal results from the

perspective of the agent. Hence, the principal would like to provide the agent enough incentives

to ensure that the agent acts in the best interests of the principal. The problem of the principal is

to determine the optimal incentive package that does not offer excessive incentives to the agent.

Furthermore, alignment of the interest between the agent and the principal cannot be

achieved at zero cost. Jensen and Meckling (1976) define agency costs to be consisting of three

components: the monitoring expenditures by the principal, the bonding expenditures by the agent,

and the residual loss from the suboptimal decisions. They argue that managerial ownership in the

firm would reduce the conflicts between management and shareholders because managers would

pay a share of the agency costs proportional to their ownership. Thus, Jensen and Meckling

(1976) suggest that management ownership is a good way to align the interests of managers and

shareholders and to reduce agency costs.

These arguments suggest that a mechanism for reducing agency problems between

managers and owners is the employment contract which specifies compensation and its

components. Shavell (1979) proposes that different forms of compensation have different

incentive effects on CEOs. Lambert, Larcker, and Verrecchia (1991) suggest that if a manager’s

compensation is tied to the stock price, the agency problem which includes overly short-sighted

behaviours can be mitigated. Gray and Cannella (1997) argue that incentive pay is related more

with long-term performance and value appreciation of firms and non-incentive compensation

8

aims at providing the CEOs with a stable stream of cash flows. Daily, Johnson, Ellstrand, and

Dalton (1998) propose that incentive pay and non-incentive pay induce different levels of risk

and incentive objectives. Arya and Huey-Lian (2004) suggest that equity-based compensation,

such as restricted shares and stock options, can serve the purpose of aligning the long run

interests of shareholders and managers. Chowdhury and Wang (2009) argue that companies are

structuring CEOs compensation as a combination of incentive pay and non-incentive pay to align

the interests of owners and CEOs.

Another mechanism for reducing agency problems between managers and owners is the

ownership structure. Fama and Jensen (1983) argue that concentrated ownership by outsiders has

the same effects as managerial ownership in reducing agency costs. Thus, the presence of a

shareholder that owns a significant portion of voting rights can be a substitute to significant

managerial ownership in large corporations, large professional partnerships, and mutual

companies. They indicate that concentrated shareholdings by outsiders create more effective

monitoring of managers, which can improve firm performance. Ang, Cole, and Lin (2002)

empirically examine how agency costs vary with a firm’s ownership structure using a sample of

1,708 small US corporations. Their results support the theories of Jensen and Meckling (1976)

and Fama and Jensen (1983) about ownership structure and the alignment of interests between

managers and shareholders. In particular, they find that when an outsider manages the firm,

agency costs are higher. Also, agency costs vary inversely with the manager’s ownership level.

Moreover, agency costs are positively related to the number of non-manager shareholders.

2.3 Empirical evidence regarding the relation between pay and performance

Section 2.2 suggests that incentive compensation is the tool to align the interests of

managers and owners. The empirical evidence confirms that incentive compensation is popular

in many countries and is widely accepted by companies. For example, Barenbaum and Schubert

(1993) find that in 1988 more than 90% of the 400 largest industrial and service companies in the

United States (US) used stock options as part of their compensation packages. However, recent

studies suggest that the relation between performance and incentive compensation is weak.

Jensen and Murphy (1990) empirically examine the relationship between executive

incentives and performance by using over 2000 CEOs data. Their results show that a 10%

change in firm value leads to 0.33% change in total CEO compensation. These results

9

demonstrate that CEO wealth is not significantly related to shareholder wealth. They argue that

CEO ownership of their firm’s stock is the largest CEO performance incentive. However, the

holdings of CEOs are small and decreasing. Leonard (1990) finds results different from Jensen

and Murphy (1990). The author indicates that corporate success does not have an impact on the

level of executive pay using 439 large US corporations over a period of 1981-1985. Kaplan

(1994) indicates that incentive pay which is tied to stock performance and to factors that are

conducive to earnings performance affects the fortunes of Japanese top executives. He also finds

that the stock performance is less related to the fortunes of Japanese managers than those of US

managers. A related study of Hall and Liebman (1998) document that firm performance is

strongly correlated to CEO compensation using data over the period 1980-1994. In particular,

they show that salary and bonus are weakly related to firm performance. However, in

comparison to salary and bonus, the equity-based pay works better to align the interests of CEOs

and shareholders. Similarly, Murphy (1999) uses a number of control variables to clarify the pay-

performance relationship. By using US data covering 1970 to 1996, and international data in

1997, he offers the following important insights. First, in larger firms, if the levels of pay are

high, the pay-performance sensitivities are low. Second, the relationship between the level of pay

and pay-performance sensitivity are more significant in industrial firms than in utilities. Third, in

the US the level of pay for performance is much higher than the pay for performance in other

countries. Fourth, although the incentive pay-performance relations are significant, managers

should not be left alone to design performance-based compensation.

More recent, many studies have shown significant positive relation between pay and

performance, but with rather weak pay-performance sensitivity (Jeppson, Smith, and Stone 2009;

Shaw and Zhang 2010; Fahlenbrach and Stulz 2011). Jeppson, Smith, and Stone (2009) examine

the relationship between CEO compensation and several measures of firm performance using

200 large public companies in 2007 which filed proxy statements with the SEC. They do not find

a strong relationship between CEO compensation and firm performance. The exception is total

revenue, but with a low R2. They also find that CEO compensation is positively related to firm

size. Similarly, Shaw and Zhang (2010) find that CEO bonus compensation is less sensitive to

poor earnings performance than it is to good earnings performance using data over the period

1992-2005. They find no evidence supporting that CEOs are punished for poor firm performance.

Indeed, CEOs are even rewarded with poor performance. Similarly, Fahlenbrach and Stulz (2011)

10

study the relation between bank performance during the 2008 bank crisis and the bonus and

equity-based compensation of bank CEOs. They find that banks with CEOs whose incentives

were better aligned with the interests of their shareholders performed worse. Their results show

that both cash bonus and stock options do not have an adverse impact on bank performance

during the crisis.

There are two studies that consider that relation between incentive pay and performance

in Canadian companies. Zhou (2000) considers executive compensation over the period 1991-

1995 inclusive and provides several insights. The results show that executive compensation is

positively correlated to firm performance but the overall relationship is weak. The author finds

that firm size has a positive impact on CEO compensation. Moreover, smaller firms exhibit a

strong negative correlation between the probability of CEO turnover and stock performance. We

extend their study by examining contingent compensation using more recent observations that

cover 5 years instead of 3 and we classify companies on the basis of their ownership structure.

Another Canadian study that may have objectives similar to those of this study is

Chowdhury and Wang (2009). Using data related to the TSE 300 firms from 1995 to 2002, they

find that contingent pay in Canada, both in monetary terms and as percentage of total pay, has

been increasing during the study period. They argue that this finding is consistent with an

implication of agency theory. Namely, boards of directors seem to be trying to raise CEO

contingent pay to ensure better performance. Their results show that firm size and investment

opportunities positively affect contingent pay. However, the key limitation in Chowdhury and

Wang (2009) is that they investigate CEO contingent compensation only in institution-controlled

firms. We extend their work by examining family-controlled as well as institution-controlled

firms.

2.4 Ownership Structure

This study examines whether ownership structure can explain the differences among the

compensation structures of chief executive officers (CEOs). Therefore, in this section we review

the findings of previous studies regarding the impact of ownership structure on corporations.

During the era of Berle and Means (1932), the theory of the firm was developed under

the assumption that organizations have widely-held ownership. La Porta et al (1999) suggest that

the widely-held ownership structure continues to be a very common form of organization in the

richest common law countries including the United States. However, Shleifer and Vishny (1986)

11

and Morck et al. (1988) show that the largest American firms have some concentration of

ownership. Shleifer and Vishny (1986) show that such concentration of shareholding can make

sense in terms of solving the agency problem. They argue that if there are many small

shareholders, each will try to take a free-ride on the issue of monitoring the managers. In that

case no monitoring would occur. Hence, if there are large shareholders they would solve the

free-rider problem. In addition, Morck et al. (1988) find that large firms, outside the United

States and the United Kingdom, normally have controlling owners, such as families. These

controlling families maintain their significant influence through various mechanisms such as

pyramidal control structures, cross-shareholdings, and super voting rights. Such mechanisms

allow the families to remain in control even without making commensurate capital investment.

Therefore, given the enormity of these corporations, such families have considerable power in

controlling significant proportions of their countries’ economies.

In a later study, La Porta et al. (1998, 1999) indicate that firms in other developed and

developing countries have a higher level of ownership concentration. The study shows that,

managerial ownership aligns managers’ and outside shareholders’ interests at low levels of

managerial ownership. Up to a certain level of managerial ownership, managers would like to

maximize the firm’s value. However, if the managerial ownership achieves and passes an

optimal level, managers focus on maximizing their own benefits, such as undertaking high-risk

projects, resisting a takeover, and building empires at the expense of other shareholders in the

firm.

The first study to examine the issue of ultimate control is that of La Porta et al. (1999).

Studying ownership structures of large firms in 27 wealthy economies, they find that if we trace

the ultimate owners, we will find relatively fewer firms with widely-held ownership. Even the

largest firms have controlling shareholders, such as families or states. This study underscores the

importance of ownership pyramids through which an ultimate owner could control other

companies by means of indirect ownership. An ultimate owner, who has the most voting rights

(instead of cash flow rights), can be found by tracing the chain of ownership. In addition, the

study shows that different kinds of ownership and control have different impacts on the wealth of

large shareholders.

The idea of ultimate ownership, which was initially propagated by La Porta et al (1999),

instigated a number of related empirical works. First, Claessens, Djankov, and Lang (2000)

12

analyze the separation of ownership and control by using data from nine East Asian countries.

They find that pyramid structures and cross-holdings improve corporate control in all studied

countries. Second, Claessens, Djankov, and Lang (2000) use 1,301 publicly traded corporations

in eight East Asian economies to disentangle the incentive and entrenchment effects of large

ownership. They find that a positive relationship exists between the cash-flow ownership of the

largest shareholder and firm value. However, when the control rights of the largest shareholder

exceed its cash-flow ownership, firm value falls. They also find that managers at family-

controlled firms have more ways to divert benefits to themselves than managers at firms with

widely-held ownership. Third, Faccio and Lang (2002) study the ultimate ownership and control

of corporations. They use data of 5,232 firms in 13 Western European countries. They find that

there are more family-controlled firms than widely-held firms. Also, most financial and large

firms are widely-held firms, while most non-financial and small firms are family-controlled

firms. Furthermore, studies show that in the US and the UK, firms are mostly characterized by

dispersed ownership. However, most of continental Europe and Asia are commonly

characterized by ownership controlled by individuals, families, governments or industrial groups

(La Porta et al 1999, Faccio and Lang 2002). Last, unlike other countries, in China, the

government controls the majority of listed firms (Kato and Long, 2006).

In terms of the nature of governance structures in Canada, Roe and Lee-Sing (1996)

observe that ownership concentration in Canada is high because individuals, families or private

holding companies are the ultimate controlling owners of many large firms. La Porta et al. (1999)

suggest that the ownership structures at Canadian firms are closely similar to those observed in

most countries. Amoako-Adu and Smith (2001) note that the existence of dual-class shares,

which is a relatively common phenomenon in Canada, facilitates concentrated ownership and

family control. Klein, Shapiro, and Young (2005) examine the Canadian evidence on the

relationships between corporate governance, family ownership, and firm value. They suggest

determining the ultimate control by using voting rights instead of equity ownership. They do not

study CEO compensation directly but they test the relationship between firm value and the newly

released indices of effective corporate governance.

2.4.1 Widely-held group versus the concentrated group

13

In our study, we examine the compensation structure of three distinct groups: family-

controlled firms, institution-controlled firms and widely-held firms. We distinguish these three

kinds of firms to represent different levels of market imperfection. Compared with family-

controlled and institution-controlled firms, widely held firms have dispersed ownership. Demsetz

(1983) mentions that since ownership is widely dispersed across many shareholders, no

shareholders can exercise real power to oversee managerial performance in modern corporations.

Shareholders, owning a low amount of shares, have little or no incentives to exert monitoring

behavior (Grossman and Hart, 1988). Thus, managers in widely-held firms have more freedom in

using firm’s capital than managers in non-widely-held firms. Shleifer and Vishny (1997) suggest

that in absence of monitoring, managers would like to maximize their own utilities instead of

shareholders. Another empirical work of Healy and Cole (2002) shows that absence of a

stockholder with a large proportion of stock increases the agency costs, leading to the use of

compensation contracts based on performance.

The most significant weakness of a widely-held ownership structure is the lack of

shareholder monitoring due to the unmatched benefit and cost of monitoring for small

shareholders. The existence of at least one large shareholder will reduce the costs of monitoring

and agency costs and asymmetric information. Major shareholders mitigate the conflict between

managers and shareholders (Shleifer and Vishny, 1986). Shleifer and Vishny (1997) empirically

examine the consequences of corporate ownership for corporate valuation using data on

companies from 27 wealthy countries around the world. They find that compared to small

shareholders, large shareholders have greater resources and incentives to monitor managers

reducing some agency costs.

In particular, large shareholders may be willing to spend time and effort to collect more

information on management performance or to estimate the firm’s investment projects and thus

reduce the information asymmetry. Theoretically, shareholders with significant shares have more

incentives to monitor management. As a result, they (large shareholders) are able to monitor

more efficiently (La Porta et al, 1999). Bebchuk and Stole (1993) suggest if an investor holds a

larger block of shares, this investor will have stronger incentives to protect the investment by

monitoring management. Firth, Fung and Rui (2006) investigate the relationship among agency

costs, ownership structure, and governance mechanisms by using data from China listed firms.

They find that the level of agency cost is not significantly related to individual shareholding,

14

institutional shareholding, and government ownership. Their results indicate that agency costs

are lower because of concentrated ownership, but they are not lower for the boards with a

majority of outside directors. Hence, they argue that though Chinese public firms are undergoing

ownership and governance reforms, such reforms have not yet led to lower agency costs. Their

results support some prior empirical results for the US firms (Singh and Davidson, 2003). A

related study of of Florackis and Ozkan (2008) shows that managerial ownership plays a

significant role for corporate governance mechanism for the UK firms over period from 1999 to

2003. Their results indicate that both compensation and ownership concentration are important

factors in mitigating agency problems. Besides, they find that executive ownership has a positive

relationship with growth opportunities, which means, more growth opportunities firms offer

more incentive mechanisms. Similarly, Ozkan (2007) empirically examines the impact using a

sample of 414 large UK companies for the fiscal year 2003/2004. They also find that institutional,

block-holder ownership, and directors’ are negatively related to CEO compensation

In addition, enhanced monitoring will decrease the waste of free cash flow by managers.

In the USA, publicly traded family-controlled firms (which constitute about one third of the total

listed firms) have higher Tobin’s q values and higher return on assets than nonfamily-controlled

firms (Anderson and Reeb, 2003). Ben-Amar and Andre (2006) state that family ownership has a

positive impact on value creation.

2.4.2 Family-controlled Group Versus Institution-controlled Group


However, whether or not the control shareholders are playing an active monitoring role is still an

important issue. Empirical research suggests that institutional investors play an important role on

firm strategies, for example, executive/CEO compensation (Smith 1996; David, Kochhar, and

Levitas, 1998). Smith (1996) concludes that when shareholder activism is successful in changing

governance structure, shareholder wealth will increase. We can argue from Smith’s study that

institutional activism does have impact on share price, which again could affect stock-based

CEO compensation. Similarly, David et al (1998) examine whether institutional investors have

an impact on CEO compensation policy or not. Their results show that institutional owners with

only an investment relationship with a firm influence compensation in accordance with

shareholders preferences to “(1) lower its level and (2) increase the proportion of long-term

15

incentives in total compensation.” Moreover, some other studies have examined the effect of

institution ownership on CEO/executive compensation (Hartzell and Starks, 2003). Hartzell and

Starks (2003) use a sample of 1,914 firms from S&P’s ExecuComp between 1992 and1997.

They find that the concentration of institutional investor ownership is positively tied to the

performance sensitivity of managerial compensation and is negatively tied to the level of that

compensation. They also find a positive relationship between institutional investors and

executive compensation. Besides, their results imply that institution-controlled firms tend to use

incentive compensation to mitigate the agency problem between shareholders and managers.

From the point of aligning the interests of managers and shareholders, the family-

controlled group is superior to the institution-controlled group. According to Jensen and

Meckling (1976), family-controlled firms should be characterized by reduced problems of

agency and agency costs. This hypothesis has been tested and confirmed by Chrisman, Chua, and

Litz (2004). They suggest that the overall agency problems in family-controlled firms is less than

that in non-family-controlled firms using 1,141 small privately held US firms. Demsetz and Lehn

(1985) show that family-controlled firms face less agency problems because they (family-

controlled firms) are able to monitor their managers directly. Similarly, Mehran (1995) examines

the relationship between executive compensation structure and ownership using 153 randomly-

selected manufacturing firms. Results show that firms with a larger percentage of their shares

controlled by outside block-holders offer less long term incentive pay, implying that block-

holder monitoring is a substitute to the incentive components of compensation. Another

empirical work of Kole (1997) indicates that the likelihood of any form of explicit compensation

arrangement is reduced by the presence of an agent of founding family on the board.

First, institutions are more flexible in switching their ownership from one firm to another

depending on performance. If the costs of monitoring are high in comparison to the costs of

rebalancing portfolios, institutions will choose to rebalance instead of monitoring. Unlike

individuals or families, institutions invest money of other people. Institutions have the obligation

to know and protect what they invest. They should take proactive actions, so that the

management of investee firms works towards maximizing shareholder value. (David, Kochhar,

and Levitas 1998). Consistent with internal monitoring of management, substantial top

management changes is negatively related to a firm’s stock returns (Warner, Watt, and Wruch,

1988). They examine this relationship using the sample consists of 269 firms listed on the New

16

York and American Stock Exchanges in the period 1963-1978. Results show that the ratio of the

number of top management changes to the number of firms is relatively stable at 0.183 for all

changes. They indicate that this relationship is consequence from monitoring by the board, other

top managers, or shareholders. Huson, Parrino, and Starks (2002) examine CEO turnover at

large public firms listed in the Forbes over a period of 1971-1994. Results show that the

frequency of forced CEO turnover and the frequency of outside succession are increased. They

also indicate that from the beginning to the end of the period, the relationship between the firm

performance and the likelihood of forced CEO turnover remain the same, even though the

internal mechanisms is significantly changed. Thus, the characteristics of internal monitoring

mechanisms and the nature of CEO turnover do not influence the sensitivity of forced turnover to

firm performance. Kaplan and Minton (2006) studies CEO turnover using data from large US

companies spanning a period of 13 years from 1992 to 2005. The authors find that CEO’s tenure

on average is less than 7 years. Compared to previous studies, the results show that the annual

CEO turnover rate has been increasing. Furthermore, the average tenure of CEOs drops to six

years by using data from 1998 to 2005. They analyze the impact of three components of firm

performance (performance relative to industry, industry performance relative to the overall

market, and the performance of the overall stock market) on internal turnover. Results show that

these three factors have stronger impact on internal turnover after 1998. In contrast, a family that

controls a firm does not have this flexibility.

Second, family-controlled firms generally assign influential positions to family members

whose focus is in line with that of the family group. Even though a non family member may be

appointed as the manager, the level of monitoring is significant given the high ownership

concentration by the family. Anderson and Reeb (2003a) state that family firms are managed or

controlled by founding families. About one-third of the S&P 500 firms across a broad range of

industries are characterized by such ownership. They also suggest that family owners have better

knowledge of the firm’s business activities. Such knowledge helps the owners in detecting

manipulations of stock or firm performance, if any. Bennedsen et al. (2007), using data from

Denmark, show that a professional CEO provides much better performance in a family firm than

a CEO who is a member of the family. Similarly, Villalonga and Amit (2004), upon their study

on all Fortune 500 firms over period 1994-2000, argue that family ownership creates value only

when a founder of the firm or non-family members serve as CEOs. Ben-Amar and Andre (2006)

17

conclude from Canadian data that separation of ownership and control does not negatively affect

value creation. They indicate that family ownership in Canada is a positive factor in value

creation. However, the level of monitoring by a family may not necessarily translate into a

reduction of agency costs for the minority shareholders. Indeed, previous studies suggest that

significant family ownership may lead to agency costs of its own. The family may divert

company resources for its own benefit despite the presence of a manager who may or may not be

a family member. Essentially, the family and the manager can collude to spend on perks and

personal benefits at the expense of minority shareholders. Schulze, Lubatkin, and Dino (2003)

suggest that agency benefits gained by family-controlled firms are offset by free-riding and other

agency problems. Chourou (2010) use a panel of Canadian companies ultimately controlled by

families over a period 2001-2004 to examine hypothesis of owner managers expropriating

minority shareholders by receiving excessive compensation. He suggests that excessive

compensation of chief executive officers at some family owned Canadian corporations may

cause expropriation of minority rights.

2.5 Other Factors

Early studies, for example Baumol (1959) and Lewellen and Huntsman (1970), examined

that factors that determine CEO compensation. They find that firm size and firm performance

have an impact on CEO compensation. Ciscel and Carroll (1980) argue that the findings of these

studies are limited by multicollinearity problems. They find that the market for managerial talent,

the external performance of the firm, and the internal technical efficiency of production affect

the level of executive compensation, after correcting for multicollinearity.

Some studies examine the relationship between CEO compensation and market or

industry performance. Hart (1983) indicates that when a manager owns a small stake the product

market may still force managers to follow the principle of shareholder value maximization.

Hart’s argument is based on the assumption that a given product has a cost component that is

common among all producers. If agency costs make the costs of a product higher than the costs

of its peers, consumers will avoid buying it. This result affects negatively the manager’s personal

benefits.

Jensen and Ruback (1983) focus on the role of the market for corporate control. They

investigate the relationship between managers and shareholders and corporate takeovers. Their

18

results show that corporate takeovers benefit the target firm’s shareholders. In addition, the

shareholders in the bidding firm do not lose. More recently, Tannous and Cheng (2007), propose

that the market for corporate control provides another incentive for managers to perform. They

provide evidence suggesting that corporate takeovers are often motivated by the poor

performance of the target and turn around plans that include dismissal of existing managers.

Gibbons and Murphy (1990) argue that corporate performance depends on non-

controllable factors such as industry and market conditions, which also have an impact on CEO

compensation. They examine the relationship between relative performance and CEO

compensation. Their empirical evidence strongly supports the existence of a positive relation

between CEO pay and firm performance but they find that CEO pay is negatively related to

industry and market performance. They also find that CEO performance is more tied to

aggregate market movements than industry movements.

Jacobs (1991) shows that CEO compensation is correlated to market effects. He argues

that managerial short-sightedness is a key reason for the decline in American business

competitiveness. Such short-sighted behaviour of managers raises questions regarding whether

better designed compensation contracts could induce managers to behave in a way that is

consistent with the long-run interests of shareholders.

19

CHAPTER 3

Theoretical Arguments and Hypotheses

Past studies indicate that the ownership structure affects the degree of agency costs. It is

argued that the higher the ownership of a particular entity the lower will be this entity’s

monitoring costs in proportion to its benefits of monitoring the managers. We propose that the

ownership structure is a spectrum that ranges from full ownership by one individual to widely

dispersed ownership by a large number of shareholders each owning a very small portion of the

firm. Furthermore, we propose that the costs of monitoring management effectively are

significant but these costs are fixed while the benefits of monitoring management are

proportional to the percentage of ownership in the firm. Thus, shareholders who have significant

ownership in a firm should be willing to monitor managers closely, which suggests that

concentrated ownership in a firm should increase the level of monitoring and reduce agency

costs.

Similarly, previous studies suggest that the compensation contract can be designed to tie

a portion of the pay to performance. It is argued that performance-based compensation will align

the interests of managers and shareholders. Therefore, it should reduce the need for shareholders

to monitor the performance of managers. Therefore, if we assume that shareholders are indeed in

a position to structure compensation packages as they please regardless of the ownership

structure then we can suggest the existence of a negative relation between the degree of

monitoring and the proportion of performance-based CEO compensation. Under this scenario the

performance-based compensation may vary depending on the ownership structure of firms. The

widely-held firms, in which ownership is widely dispersed across many shareholders, should rely

heavily on performance-based compensation to align the interests of managers and shareholders

and reduce agency costs. In contrast, firms which have concentrated ownership should have less

need for performance-based compensation. In these firms, the existence of at least one

shareholder with a significant ownership stake will improve monitoring and reduce the level of

agency costs.

However, there is no evidence that suggests shareholders are in control of compensation

packages. In the contrary, the evidence suggests that executive compensation is mainly

determined by competitive pressures in the market for CEOs and by benchmarking. We propose

that in this environment concentrated control by an institution or a family may or may not lead to

20

significant differences in incentive compensation across ownership structures. However, in the

presence of high monitoring by families or by institutions we should find a significant relation

between incentive compensation and performance.

Previous studies suggest that performance based compensation can take many shapes and

forms. The annual compensation of a CEO in a typical firm has three components that account

for approximately 90% or more of the CEO’s total compensation.1

The three most significant

components are fixed annual salary, bonus, and contingent compensation (contingent

compensation consists of stock options, performance plans, restricted options, and other long-

term incentives). The bonus is usually based on some accounting metric such as return on assets,

return on equity, or cash flow per share. Theoretically, these metrics are positively related to the

value of the firm and good performance along these metrics will improve shareholder’s value.

Furthermore, it is easy to link performance in accounting measures to managerial actions.

Therefore, it is preferable in environments of active monitoring. However, the bonus as an

incentive pay may be criticized on the basis that it focuses the attention of managers on short-

term performance and distracts from capitalizing on the long-term interests of shareholders.

Stock-based compensation is introduced to align the long-term interests of shareholders with the

interests of managers. It is argued that managerial equity ownership in a firm leads CEOs to

manage in the best interests of shareholders because these interests are their own interests.

However, there are wide differences of opinion among academics, practitioners, and policy

makers regarding the effectiveness of stock-based compensation in promoting the long-term

interests of shareholders. First, stock-based compensation may reward managers for success

resulting from factors beyond the control of managers. Second, managers may not hold enough

cash flow interest to align their interests with those of the owners. Third, there is no universally

accepted model or method that can help owners to set the optimal stock-based compensation

component to prevent excessive compensation. Therefore, we propose that shareholders who are

in control of compensation packages are likely to prefer bonus compensation as a way to control

the actions of managers and they are likely to offer equity based compensation based on market

practices benchmarking methods.

1 The remaining portion consists of annual pension, annual vocation, lump-sum pension, and retirement

allowances.

21

In summary, our hypotheses are as follows:

H1a: The proportion of CEO annual bonuses in concentrated firms is lower than the

proportion of CEO annual bonuses in widely-held firms.

H1b: The proportion of CEO contingent compensation in concentrated firms is lower than the

proportion of CEO contingent compensation in widely-held firms.

Concentrated ownership may be the result of high percentage ownership by a family or

by an institution. We propose that either form of control should improve monitoring and reduce

agency costs but if all else are equal control by a family provides better monitoring than control

by an institution. Therefore, our hypotheses can be expanded to include:

H2a: The proportion of CEO annual bonuses in family-controlled firms is the largest among

the three types of ownerships: family-controlled, institution-controlled, and widely-held.

H2b: The proportion of CEO annual bonuses in institution-controlled firms is lower than that

in family-controlled firms and is higher than that in widely-held firms.

H2c: The proportion of CEO annual bonuses in widely-held firms is the lowest among the

three types of ownerships: family-controlled, institution-controlled, and widely-held.

H3a: The proportion of CEO contingent compensation in family-controlled firms is the lowest

among the three types of ownerships: family-controlled, institution-controlled, and

widely-held.

H3b: The proportion of CEO contingent compensation in institution-controlled firms is higher

than that in family-controlled firms and is lower than that in widely-held firms.

H3c: The proportion of CEO contingent compensation in widely-held firms is the highest

among the three types of ownerships: family-controlled, institution-controlled, and

widely-held.

Previous empirical studies on CEO compensation concentrate mainly on the pay to

performance relationship. Earlier studies, for example Jensen and Murphy (1990), Kaplan (1994),

and Hall and Liebman (1998), argue that incentive pay is related to firm performance. However,

22

recent studies suggest that the relation between performance and incentive pay is weak. Shaw

and Zhang (2010) find that CEO bonus compensation is less sensitive to poor earnings

performance than it is to good earnings performance. Similarly, Fahlenbrach and Stulz (2011)

find no evidence to support the proposition that the performance of banks during the 2008

financial market crisis is positively related to the alignment of incentives of bank managers and

their shareholders. We examine the relation between incentive pay and firm performance to see

whether controlling for ownership structure can clarify this relation.

H4a: Annual bonus of CEOs is positively related to firm performance and this relation varies

across ownership structures

H4b: Contingent compensation of CEOs is positively related to firm performance and this

relation varies across ownership structures

We examine how incoming CEOs are compensated in comparison to their predecessors.

We propose that experience in the industry and in similar position may enable the incoming CEO

to negotiate a high compensation package and a structure that is in the best interests of the CEO.

On the other hand, the departure of a CEO could be seen as an opportunity for a firm to re-

establish its own priorities and to design the compensation package to promote the interests of

shareholders and the ultimate power brokers of the firm. Accordingly, we expect that the

compensation of the incoming CEOs would be structured differently than the compensation of

their predecessors. Furthermore, we propose that the structure of the compensation packages of

incoming CEOs will vary depending on the ownership structure.

H5a: The bonus compensation of incoming CEOs is lower than the bonus compensation of

their predecessors. H5b: The contingent compensation of incoming CEOs is higher than the contingent

compensation of their predecessors.

23

CHAPTER 4

Data

4.1 Ownership Structure

The sample selection starts by considering the 269 firms that made up the S&P/TSX

Composite Index (formerly known as the TSX 300) at the end of 2007. These firms comprise

approximately 71% of the market capitalization of all Canadian-based companies listed on the

TSX. Our data spans the years 2003 to 2007 inclusive. Because the components of the S&P/TSX

index vary from year to year, we choose the components of the 2007 index as the initial sample

and then we track back through the sample period to select only the firms that have continuous

presence on the S&P/TSX index throughout the sample period. By doing this, we get 143 firms

which have full records.2

Data on ownership structure are manually collected from the Inter-Corporate Ownership

(ICO) database which is released quarterly by Statistics Canada.

3

From the ICO, we find that there are three overwhelming kinds of ownership structures:

family-controlled, institution-controlled, and widely-held. The widely-held group consists of all

companies that do not have a controlling interest of 10% or more. We define a firm as family-

controlled if its controlling shareholders who own 10% or more are from a family-controlled

group. Institution-controlled are firms with controlling shareholders from institutions such as

pension plans, mutual funds, trusts, banks, and insurance companies. The categorization between

widely-held, institution-controlled, or family-controlled is done each year from 2003 to 2007

inclusive which allows ownership to change over time. We observe that compared to the other

two groups, the family-controlled firms have the most stable ownership structure over time.

We choose the last quarter of

each year to represent the firms’ ownership for the entire year. We rely on the annual reports

instead of the quarterly reports because the quarterly reports contain missing data. The

accounting data is collected from the Compustat database.

In our study, a firm is categorized as having controlling shareholders if 10% of the voting

shares of the firm are owned by a single individual, a group of individuals acting together, a

2 Our data sample may be criticized on the basis that the method by which we selected the firms will subject the

results to survivorship bias. However, our main objective is to compare the incentive compensation among the three ownership groups. We expect that the survivorship bias, if any, will have the same impact on each of the three groups. Thus, the survivorship bias will not affect the comparison results.

3 Statistics Canada requires publicly held companies to identify any controlling interest of 10% or more. Failure to comply with this requirement violates Canadian laws and subjects the violators to penalties.

24

family, an institution, or another corporation. Previous studies, for example La Porta et al. (1999),

show that this threshold is sufficient to control a firm’s decision making system. Furthermore,

other Canadian databases, such as the Financial Post (FP) Informat, use this threshold to classify

firms between concentrated ownership and dispersed ownership.

While the 10% cut off is sufficient, it may not be necessary. Critics may argue that the 10%

cut off is too high as ownership interests less than 10% may be effective in controlling a

company. We agree with this argument but we cannot lower the cut off ownership level. The

ICO database from which we obtain the control information does not report on ownership

interests less than 10%. We decided to rely on the ICO despite its limitations for two reasons.

First, we feel that it is the most reliable source given that the information is collected to comply

with government regulations. Second, we considered the possible effects of using a high cut off

level and concluded that it is not affecting the qualitative results. In our view, using a 10% cut

off level instead of a lower one, for example 5%, would improperly classify some firms as

widely-held when they should be classified as family-controlled or institution-controlled. The

impact of this misclassification would be to weaken the differences between the widely-held

firms and each of the other two groups. Our results suggest that the differences are significant

despite the possible misclassifications.

According to La Porta et al (1999), an ultimate owner is the entity that has the most

voting rights instead of cash flow rights. For example, a family-controlled firm is controlled by

an individual or family owning 10% or more of the firm’s voting stock. For this purpose, we

track the ownership structures of the direct controllers and categorize firms accordingly. For

example, Ensign Resource Service Group Inc. is controlled by the Mackenzie Financial

Corporation through 12.33% of the voting rights. Meanwhile, Mackenzie Financial Corporation

is 100% controlled by a family-controlled group. Under this case, Ensign Resource Service

Group Inc. is categorized as family-controlled firm. In our sample, there are several companies

for which the direct controllers are institutions while the ultimate controllers are families.

4.2 CEO compensation

CEO’s compensation data are hand-collected from the proxy circulars of each company

as listed in the System of Electronic Document Analysis and Retrieval (SEDAR)4

4 SEDAR is a comprehensive, online archive of securities documents filed by publicly traded companies in Canada.

. During our

25

study period, the majority of sample firms (87) had the same Chief Executive Officer over the

entire 5-year period. We call this subsample the Permanent CEO Group. The remaining 39 firms

experienced one or more changes in CEOs during the study period. We call this subsample the

Transient CEO Group. Since 8 firms changed ownership structures in the transient CEO group,

we use 31 (39 minus 8) firms to compare the compensation of the new CEOs with the

compensation of their predecessors.5

We combine the Permanent CEO Group and the Transient CEO Group to conduct

multivariate analysis. For the 87 firms that make up the Permanent CEO Group, we obtain the

market-to-book (M/B) ratio, the return on assets (ROA), and the debt-to-equity (D/E) ratio of

each company from the Compustat database. For 8 firms, ratios are not reported. Thus, for the

subsample of permanent CEO firms, we have 79 firms. In the Transient CEO Group, ratios of 6

firms are not available. Therefore, for the subsample of transient CEO firms, we have 33 firms.

As the study covers the period 2003-2007 inclusive, for this group we have a total of 560

observations. The composition of the observations is as follows: 96 from family-controlled firms,

228 from institution-controlled firms, and 236 from widely-held firms.

Eight firms changed CEOs more than once during the 5-

year period of our study.

The components of CEO compensation vary among various companies. In this study, the

analysis of CEO compensation is based on five different measures of compensation: total

compensation, salary, annual bonus, contingent compensation, and all other compensation.

Salary measures the component of compensation that is fixed at the beginning of the year. The

annual bonus is the short-term incentive which is often based on accounting measures of

performance. Contingent compensation consists of stock options, performance plans, restricted

options, and other long-term incentives6

5 The compensation of departing and the incoming CEOs are reported separately in the compensation information

we found in the proxy circulars. If firms report compensation for a portion of the year for the departing and the incoming CEOs, we annualize the compensation.

. The term ‘contingent’ indicates that such compensation

depends on the performance of the underlying asset. All other compensation consists of annual

6 Some firms specify the maximum, target, and threshold of future payouts A firm grants a number of shares to its CEO, and it sets performance goals for the CEO. If firm performance is superior, the CEO can get a contain percentage (e.g.120%) of grant shares, which is the Maximum. If firm performance is average, the CEO gets the same number of shares as they are granted, which is the target. If firm performance is below the required level, the CEO will only get a percentage less than the target (e.g. 80%), which is threshold. We use the target to estimate the compensation.

26

pension, annual vocation, lump-sum pension, and retirement allowances. Total compensation is

the sum of salary, annual bonus, contingent compensation, and all other compensation.

As salary, annual bonus and all other compensation are typically paid in cash or cash

equivalent, the valuation of these three components is straightforward. However, the valuation of

the contingent component is complex. We use the Black-Scholes Option Pricing model to

calculate the dollar value of total contingent compensation.7

Furthermore, to be consistent with prior studies (Jensen and Murphy 1990a and Zhou

2000), we use the standard deviation of the continuously compounded monthly return as our

volatility. We use the monthly total return index over a three-year period ending with the grant

date. We obtain this data from DataStream. This approach may be criticized on the basis that

weekly and annual return may be more appropriate. First, weekly returns have higher

autocorrelations than monthly and annual returns. Second, using annual returns to calculate

volatility requires data from prior years and that may reduce the number of companies in our

sample due to missing data. Third, the volatility as determined from prior years, many of which

may be far from the grant date, may not reflect the true volatility of the underlying security.

This process involves several steps.

First, we need to know the number of options granted the exercise price, and the time to expiry.

These data are collected from the proxy statements. Second, we use the 3-month Treasury bill

rate as the risk-free rate. We obtain this data from the Bank of Canada. Third, we assume the

time period is monthly. Then, if companies provide the grant date only for the grant year, but do

not provide a grant date in later years, we assume that the grant date in later years is the same

date as the grant year. Fourth, if companies do not give the grant exercise price, we assume that

the closing share price on the grant date is the grant exercise price.

Moreover, a number of firms report CEO compensation annually in US dollars. Some of

these firms also report the average exchange rate for the year. For these firms, we convert the

CEO compensation from US dollars to Canadian dollars using the reported exchange rate. If

firms reported CEO compensation annually in US dollars, but do not report the exchange rate,

we use the average exchange rate for the year as reported by the Bank of Canada. In addition,

7 Canadian firms under the TSX do not need to report the values of option grants. We estimate the monetary

values of options by the Black-Scholes formula (Fischer Black and Myron Scholes 1973). Option value 𝐶0 = 𝑆0 × 𝑁(𝑑1) − 𝐸/(1 + 𝑅𝑓)𝑡 × 𝑁(𝑑2), where 𝑑1 = �ln(𝑆0 𝐸⁄ ) + �𝑅𝑓 + 1/2 × 𝜎2� × 𝑡�/�𝜎 × √𝑡�, and 𝑑2 = 𝑑1 − 𝜎 × √𝑡. In the formula, 𝑆0 is the value of the stock at the date, 𝐸 is exercise price, 𝑡 is expiration term (in months), 𝑅𝑓 is the risk-free interest rate on 3-month Canadian Treasury Bills, 𝜎 is volatility, and N(•) is the cumulative standard normal distribution function.

27

with the intention of comparing compensation of departing and incoming CEOs, we record

salary, bonus, and contingent pay in the year of the grant for both departing and incoming CEOs

in the Transient group.

4.3 Variables

4.3.1. Dependent variables

We analyze three measures of CEO compensation: annual bonus, contingent pay, and

total pay. Contingent pay includes securities under options (SUO), stock appreciation rights

(SAR), value of restricted share units (LTIP), and contingent incentive pay. Total pay includes

salary, annual bonus, contingent pay, and other pay. In addition, we use the percentage of each

category of compensation (out of total compensation) as our dependent variable. In addition, we

use the compensation per dollar of total assets as our dependent variable.

4.3.2. Control variables

Our review of prior research suggests the inclusion of four control variables in our

analysis. Total assets represent one measure of size. This measure was used by Daily et al. (1998)

and Chowdhury and Wang (2009). Chowdhury and Wang (2009) found a positive relationship

between total assets and executive compensation in Canadian corporations. Therefore, in our

original model, we control for firm size by using the natural logarithm of assets.

We use Tobin’s q ratio to control for the growth opportunities of the firm. Tobin’s q

measures the market value of a firm’s assets in relation to their replacement cost. Many studies,

for example Hartzell and Starks (2003) and Harvey and Shrieves (2001), document a strong

relationship between growth opportunities and the presence of incentive compensation. In

addition, Chowdhury and Wang (2009) show that growth opportunities are positively related to

incentive compensation of institution-controlled firms in Canada. We use Tobin’s q to account

for the variation in firm performance and to control for the presence of growth opportunities. We

use the market-to-book (M/B) ratio as a proxy for Tobin’s q. We download the appropriate data

from the Compustat database.

Consistent with David et al (1998) and Chowdhury and Wang (2009), we use the return

on assets (ROA) to measure firm performance. Chowdhury and Wang (2009) did not find a

relationship between financial performance and executive compensation of Canadian firms while

28

Zhou (2000) suggests that an overall weak relationship exists between executive pay and

performance in Canadian firms.

We use the debt to equity ratio as one of the explanatory variables. Previous studies, for

example Healy and Cole (2000), argue that CEOs of firms with a high level of indebtedness

would prefer less contingent pay to avoid the increased risk, because the amount of cash

available for either dividends or cash compensation is influenced by debt. However, owners

would like to offer CEO more stock-based compensation to prevent the CEO from choosing a

debt to total equity ratio that is suboptimal to the stockholders. Hence, we expect a positive

relation between the debt to equity ratio and incentive pay.

4.3.3. Dummy variables

Our main concern objective is to examine the relation between the ownership structure

and CEO compensation. We start by comparing CEO compensation at the concentrated group

with CEO compensation at the widely-held group. Dcon is the dummy variable that takes the

value of 1 if a firm belongs to the concentrated ownership group and 0 otherwise. Then, the

sample is divided between widely-held, institution-controlled, and family-controlled with the

widely-held group serving as the base group. DF is the dummy variable that takes the value of 1

if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value

of 1 if a firm is institution-controlled and 0 otherwise. After that, we switch the base group from

widely-held group to institution-controlled group, while DW is a dummy variable that takes the

value of 1 if a firm is widely-held group and 0 otherwise.

In addition, we examine the industry effects. Given the limited sample size, we divide our

sample into four industries based on the list of companies in the TSX Sector Indices: S&P/TSE

Canadian Energy Sector Index, S&P/TSE Canadian Financials Sector Index, S&P/TSE Canadian

Materials Sector Index, and other Indices.8

8 The other S&P/TSX sector indices are: Canadian Consumer Discretionary Canadian Consumer Staples, Canadian Diversified Metals & Mining, Canadian Gold Index, Canadian Health Care, Canadian Industrials, Canadian Information Technology, Canadian Real Estate, Canadian Telecommunication Services Sector, and Canadian Utilities

We use other indices as the base index and add three

dummies to represent the financial industry (DFin), the energy industry (DEgy), and the material

industry (DMat). DFin is a dummy variable that takes the value of 1 if a firm is a member of the

financial industry and 0 otherwise. DEgy and DMat are defined in the same manner.

29

We use another dummy variable, D1YC, to control for newly appointed CEOs. D1YC

takes the value of 1 if the CEO is in her/his first year on the job and 0 otherwise.

Finally, we control for the year effects. We use year 2003 as the base year and add the

four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005,

2006, and 2007.

30

CHAPTER 5 Descriptive Statistics and Univariate Tests

In this chapter, we examine the trends of CEO monetary compensation over the period

2003-2007. For this analysis, the focus is on the compensation of CEOs who remained in the

same position throughout the study period. Each component of CEO compensation is calculated

as a percentage of total pay. In addition, for transient CEO firms we compare CEO compensation

before and after the turnover. The analysis is conducted for the three different ownership

structures: family-controlled, institution-controlled, and widely-held. Finally, for both permanent

CEO firms and transient CEO firms we present descriptive statistics and univariate tests to

compare the family-controlled, institution-controlled, and widely-held firms.

5.1. Permanent CEO firms

The following tables and figures analyze CEO compensation across different years. Table

1.1 presents the salary, annual bonus, contingent pay, and total compensation paid by permanent

CEO firms. The table shows that all components of compensation have been increasing in

Canada over the five-year period of 2003-07. We find that the family-controlled group has the

highest pay in terms of salary, annual bonus, contingent pay, and total compensation. In addition,

the growth rate of the CEO compensation is higher in the family-controlled group than in the

other two groups. In 2003, the institution-controlled group has the lowest pay among the three

groups. All components of CEO compensation have been increasing steadily from 2003 to 2006.

This finding is consistent with the conclusions of previous studies which show that the

contingent component of CEO compensation has increased in Canada (Zhou 2000; Chowdhury

and Wang 2009). From 2006 and onwards, the components of compensation in the institution-

controlled group were higher than the components of compensation in the widely-held group. In

widely-held group, we see that the salary and annual bonus have increased sharply, but there is

no change in the contingent pay. Thus, in 2007, the widely-held group has the lowest pay among

the three groups.

[Insert Table 1.1 Here]

These findings can be observed in Figures 1.1, 1.2, and 1.3. Figure 1.1 shows that salary9

9 We obtain the annual rate of inflation from the Bank of Canada website.

,

adjusted for inflation, has been continually increasing from 2003 to 2007 for every ownership

31

group. Compared to the institution group and the widely-held group, the family-controlled group

offers the highest salary for the CEO. Figure 1.2 shows that the annual bonuses of the widely-

held and institution-controlled firms have moved higher slowly and are almost equal in 2007. In

contrast, the annual bonuses of the family-controlled group have increased at a faster rate

steadily over the years. Figure 1.3 shows that contingent compensation of the widely-held firms

is almost stable during our study period. In contrast, the contingent pay of the family firms has

increased sharply between 2003 and 2005 and stabilized in 2006 and 2007. Similarly, contingent

pay of the institution-controlled group increased steadily between 2003 and 2006 but declined in

2007. Figure 1.4 illustrates that the total pay of both the family-controlled and the institution-

controlled groups has an upward trend, whereas total pay of the widely-held firms is almost

constant.

[Insert Figures 1.1, 1.2, 1.3, and 1.4 about Here]

Table 1.2 presents each of the components of CEO compensation as a proportion of total

pay. The table shows that salary makes up a small proportion of total compensation. In 2003, the

proportion of salary to total pay was 16.53%, 18.83%, and 16.45% respectively for the family-

controlled, institution-controlled, and widely-held firms. Between 2004 and 2007 inclusive the

ratio declines for the first two groups and remains almost stable for the widely-held firms.


The table also shows that the ratio of CEO annual bonus to total pay in 2003 was almost

equal to the ratio of salary to total pay. In 2003, bonus compensation was 19.37%, 17.13%, and

18.01% of total pay respectively for the family-controlled, institution-controlled, and widely-

held firms. However, contrary to salary the proportion of bonus pay increased on average after

2003. In 2007, bonus payments account for 31.60%, 20.07%, and 24.46% of total compensation

in the family-controlled, institution-controlled, and widely-held firms. In particular, the

proportion of bonus payments in the family-controlled group experiences a sharp rise from 19.37%

in 2003 to 31.60% in 2007. In the institution-controlled group the ratio in 2003 was 18.01% and

it increased by only 2.8% over the five-year period. In the widely-held group, bonus

compensation moves up to 24.46% over the five-year period. Hence, we can conclude that the

annual bonus payments as proportion of total compensation have increased from 2003 to 2007

regardless of the ownership structure to which a firm belongs.

32

Furthermore, Table 1.2 shows that contingent pay makes up a significant proportion of

total CEO compensation. In 2003, contingent compensation accounted for 58.55%, 57.81%, and

63.95% of total pay respectively for the family-controlled, institution-controlled, and widely-

held firms. These proportions fluctuate slightly between 2003 and 2007 but contingent

compensation remained well above 50% of total pay throughout the 5-year period. For example,

the proportion of CEO contingent compensation in the family-controlled group increases from

58.55% in 2003 to 60.19% in 2005, but thereafter decreases to around 53%. Similarly, the ratio

for the institution-controlled group rises to 68.15% in 2006, but it decreases to 60.94% in 2007.

In the widely-held group, the proportion of contingent compensation in total pay decreases

steadily from 63.95% in 2003 to 52.38% in 2006 but rises to 55.11% in 2007.

Figures 2.1, 2.2, and 2.3 present each of the components of CEO compensation as a

proportion of total pay. Figure 2.1 illustrates that although S-to-TP in the institution-controlled

group is the highest in 2003, it decreased continually from 2003 to 2006. The S-to-TP has

become lower than the ratio in the widely-held group since 2006. S-to-TP in the widely-held

group is relatively stable. In the family-controlled group, the S-to-TP deceases sharply in 2003.

However, the ratio remains unchanged from 2004 and onwards.

Figure 2.2 shows the annual bonus as a proportion of total compensation. We see that the

family-controlled group offers the highest proportion for the CEO across the three groups. Also,

we see an upward trend of AB-to-TP over the five-year period. In widely-held group, AB-to-TP

moves downward from 2006. In the institution-controlled group, we observe that AB-to-TP rises

sharply from 2003 to 2004, decreases the next two years and finally recovers back in 2006.

Figure 2.3 shows that CP-to-TP follows a similar trend in the family-controlled group and

the widely-held group over the study period. In particular, CP-to-TP follows a constant pattern

from 2003 to 2005, a sharp decline for some time period and remains unchanged afterwards. In

the institution-controlled group, CP-to-TP has increased dramatically from 2005 to 2006, and

then starts to fall in 2006.

[Insert Figures 2.1, 2.2, and 2.3 about Here]

5.2 Transient CEO Group

A total of 39 firms (31% of all consistent ownership structure firms) replace their CEOs

within the 5-year period. This result implies a replacement rate of 6% of CEOs per year. This

33

turnover is consistent with other studies that look at CEO turnover around this time frame by

using US data (for example, Kaplan and Minton, 2006; Chhaochharia and Grinstein, 2009).

Since 8 firms change both ownership structures and CEOs, 31 transient CEO firms are used to

compare the compensation packages of new CEOs with the packages of their predecessors. In

the multivariate tests, we include these 8 firms in the Transient CEO Group.

Regarding the corporations that change CEOs, we treat the change event for every CEO

as a way to divide the data into two sub-samples. In the case of two changes, we have three

periods. One period before the first change, another period after the second change, and the third

period is in the middle. We give each company a maximum of one change. In the case of two or

more CEO changes, the compensation of the first departing CEO is added to the data before the

change and the compensation of the last incoming CEO is added to the data after the change.

That is, we compare the last period with the first and ignore the periods in the middle.

We find that in the family-controlled group, firms offer the new CEO contingent

incentive plans at the beginning of the term. As a result, contingent compensation and total

compensation of the first year are much higher than other years.

Figure 3.1 presents the annual salary, annual bonus, contingent pay, other compensation,

and total compensation of new and departing CEOs of the family-controlled firms. Total

compensation of the departing CEOs includes payments related to retirement. The figure shows

that the new CEOs in the family-controlled group get, on average, higher compensation in terms

of contingent compensation and total compensation. This observation is consistent with prior

studies. Chowdhury and Wang (2009) suggest that the fortunes of many Canadian companies

have climbed during the past few years because of the increasing price of oil. This could be one

of the reasons why compensation of the new CEOs is higher than the compensation of the old

CEOs.

[Insert Figure 3.1 about Here]

Figure 3.2 illustrates the annual salary, annual bonus, contingent pay, other compensation,

and total compensation of new and departing CEOs of the institution-controlled firms. It shows

that in the institution-controlled firms, the new CEOs receive almost the same salary but higher

bonus and higher contingent compensation than their outgoing peers. As a result, total

compensation is higher as well.


34

Figure 3.3 depicts the annual salary, annual bonus, contingent pay, other compensation,

and total compensation of new and departing CEOs of the widely-held firms. It shows that in

these firms, the new CEOs receive almost the same salary but lower bonus and lower contingent

compensation than their outgoing peers. As a result, total compensation is lower as well.


Our sample contains five firms whose CEOs retired during our study period. Specifically,

four of these CEOs retire while employed at institution-controlled firms and one CEO retires

while employed at a family-controlled firm. Figure 4.1 reports the same information as in Figure

3.1 after deleting the family-controlled firm whose CEO was replaced due to retirement. The

table shows that the result shown in 3.1 remains unchanged. One exception is that the new

CEOs get more other pay than old CEOs. Figure 4.2 reports the same information as in Figure

3.2 after deleting the four institution-controlled firms who’s CEOs were replaced due to

retirement. The figure shows that, on average, incoming CEOs receive higher annual bonus,

contingent pay, and total compensation than their predecessors. At the same time, incoming

CEOs are paid similar salary as the outgoing CEOs while other compensation of the incoming

CEOs is less than other compensation of the outgoing CEOs.

[Insert Figures 4.1 and 4.2 Here]

Table 2.1 presents descriptive statistics for all components of compensation (annual

salaries, annual bonuses, contingent compensation, and total pay) reported for the departing

CEOs and for the incoming CEOs. On average, the salary of departing CEOs is higher than the

salary of incoming CEOs. However, all other components of compensation are higher for the

incoming CEOs than those for the departing CEOs. Total compensation of the incoming CEOs is

also higher than the total compensation of the departing CEOs.


Panel A of Table 2.2 shows the t-test results of annual bonus and contingent

compensation in the Transient CEO Group. The evidence shows that the compensation, in terms

of bonus, contingent pay, and total pay, of incoming CEOs is not significantly different from the

compensation of outgoing CEOs.


The table also presents the results of comparing the salary, bonus, and contingent pay as

percentages of total pay of incoming and departing CEOs. We find that the contingent pay as a

35

percentage of total compensation of incoming CEOs is significantly higher than contingent pay

as a percentage of total compensation of outgoing CEOs.

In conclusion, we find that the contingent pay of incoming CEOs is higher than the

contingent pay of their predecessors particularly in the family-controlled and the institution-

controlled firms. In contrast, in the widely-held firms contingent pay of new CEOs is lower than

the contingent pay of their predecessors. This finding is consistent with Hypothesis H5b.

5.3 Aggregate Sample

5.3.1 Descriptive Statistics

Table 3 presents the summary statistics of the aggregate sample. We examine both the

dependent and control variables. Our results show that contingent pay has a higher proportion of

total pay than that of annual bonus. Also, some firms do not offer their CEO contingent pay in

some specific years. The mean of the natural logarithm of firm size is 21.77 indicating that firm

size is on average 2.848 billion dollars in assets. We use the variable return on assets (ROA) to

measure the firm’s performance. Some firms report negative ROA which means they do not

perform well during our study period. The mean ROA is 4.17%. Market-to-book ratio (M/B) is

used as a proxy for the firm’s investment opportunities. The average of M/B is 2.73. Firms with

above average M/B indicate more investment opportunities. We expect that the proportion of

CEO contingent pay in such firms would be higher than firms with less investment opportunities.

The debt-to-equity (D/E) ratio can have an impact on the availability of cash. We find that the

mean of the D/E ratio is 85.77%.10

[Insert Table 3 Here]

Table 4 shows the summary statistics of both the dependent and control variables for each

of the three ownership structure groups. Several observations can be made. We find that the role

of different components of CEO pay tends to be different across firm size. As firms become

larger, all components of compensation become higher. First, the family-controlled group has the

largest average firm size and the highest compensation pay. Second, the family-controlled group

has the highest average D/E ratio. Third, the family-controlled group has the lowest ROA and

lowest M/B ratio. Fourth, the institution-controlled and widely-held firms have similar average

firm size, average D/E ratio, average ROA, and average M/B. Fifth, the widely-held firms and

10 Quebecor World Inc has an unusual D/E ratio of 1383.31%. We ran regressions after excluding this observation from the data. The results remain unchanged. Our reported results are based on the sample that includes Quebecor.

36

the institution-controlled firms pay similar compensation in the form of annual bonus, contingent

pay, and total pay.


We separate our sample into four different industries and Table 5 presents the summary

statistics of both dependent and control variables for each industry. There are 90 observations

that belong to the energy industry, 90 observations that belong to the financial industry, 110

observations that belong to the materials industry, and 270 observations in other industries.

Energy, material, and financial industries are important industries in Canada, constituting 16%,

16% and 20%, respectively, of the total sample. Several observations can be made from the table.

First, firms in the financial industry have the largest average size while the average firm size in

the energy and material industries and other industries have similar average firm sizes. Second,

the executives in the financial services industry earn higher pay than all other industries. In

particular, financial CEOs receive notably higher pay in terms of the annual bonus and

contingent pay. Third, firms in the energy and material industries have relatively lower pay than

firms in other industries. Fourth, the financial industry has the lowest average ROA (2.34) while

the ROA of the energy and material industries is almost three times higher. The ROA in the other

firms is approximately 3.34. Fifth, the material industry has the highest average M/B of 3.06

followed by other industries (2.7) and the energy industry (2.6) while the financial industry has

the lowest M/B of 2.5. Sixth, the material industry has the lowest average D/E ratio while the

financial industry has the highest average D/E ratio.


5.3.2 Univariate Tests

We conduct t-tests to determine the significance of the differences in the compensation

levels paid by the family-controlled, institution-controlled, and widely-held firms. In particular,

we conduct t-tests to compare bonus, contingent, and total compensation. The results are

reported in Table 6.1. The table shows that both annual bonus and contingent pay of the family-

controlled firms are significantly different from their counterparts at the institution-controlled

firms or the widely-held firms. The family-controlled firms and institution-controlled firms seem

to be using different levels of incentive payments to motivate their CEOs although both have

high degree of concentration of ownership. Furthermore, Table 6.1 shows that the differences in

37

the components of compensation between the institution-controlled group and the widely-held

group are not statistically significant.


The results reported in Table 6.1 may be biased by size. We conduct z-tests to determine

the significance of the differences in the structure of compensation paid by the family-controlled,

institution-controlled, and widely-held firms. In particular, we conduct z-tests to compare bonus

and contingent pay as percentages of total compensation. The results are reported in Table 6.2.

The table shows that the family-controlled firms pay the highest bonus per dollar of

compensation ( 23.6%) while the institution-controlled firms pay the lowest bonus per dollar of

total compensation (21.6%). The widely-held group has a slightly higher percentage (21.9%)

than the institution-controlled group. However, the differences among the ratios are statistically

insignificant.


Table 6.2 also shows that the family-controlled and the institution-controlled firms pay

higher contingent compensation as percentage of total compensation than the widely-held firms.

The institution-controlled firms pay the highest proportion of compensation in the form of

contingent pay (46.8%). The family-controlled firms pay a lower fraction (45.6%). However, the

differences among the ratios are statistically insignificant.

Another observation that can be learned from Table 6.2 is that contingent pay accounts

for the highest proportion of total compensation and it ranges between 42.9% at the widely-held

firms and 46.8% at the institution-controlled firms. This observation suggests that contingent

compensation have become more and more significant in executive compensation since 1990.

Using TSX data over the period of 1993 to 1995, Zhou (2000) shows that the mean of stock

option related compensation is as high as total cash compensation. Using a Canadian sample

over the period from 1996 to 2002, Chowdhury and Wang (2009) find that the average

percentage of contingent pay to total pay in institution-controlled firms is 50.30%. Therefore, our

results suggest that contingent compensation as a proportion of total compensation have

increased during the 1996-2002 period and then decreased after 2002. It is possible that these

changes are related to the stock market performance during the period of 1996-2000.

We also investigate how the control variables including total assets, ROA, M/B, and D/E

vary among the family-controlled, institution controlled, and widely-held firms. The results are

38

shown in Table 7 from which several observations can be obtained. First, the family-controlled

firms are significantly larger in size and they have significantly higher leverage than the firms in

the institution-controlled or the widely-held firms. However, the sizes of the firms in the last two

groups are not significantly different while widely-held firms are significantly more leveraged

than the institution-controlled firms. In contrast, we find the ROA and M/B of the family-

controlled group are significantly lower than their counterparts of the other two groups. The

institution-controlled firms have similar ROA and M/B as the widely-held firms.


We analyze the correlations among the different variables in Tables 8.1, 8.2, 8.3, and 8.4.

The correlations among the different variables do not seem to be a problem in our study.





In summary, using unvariate tests we find that both annual bonus and contingent pay are

different among the family-controlled, institution-controlled, and widely-held firms. This means

that the structure of compensation of a firm’s CEO depends on the firm’s ownership structure.

Next, we use multivariate analysis to examine how ownership structure affects CEO

compensation after controlling for other factors that may influence CEO compensation.

39

CHAPTER 6

Multivariate Analysis

Previous empirical studies on CEO compensation mainly concentrate on the relation

between pay and performance. The focus of this thesis is slightly different. We examine the

effect of corporate ownership structure on the levels and structures of the CEO compensation.

Since there is no precedent research on this topic, we borrow and extend the models on pay to

performance relationship to serve our purpose.

Bertrand and Mullaninathan (2001) use the following equation to analyze agency

framework.

𝑦𝑖𝑡 = 𝛽 × 𝑝𝑒𝑟𝑓𝑖𝑡 + 𝛾𝑖 + 𝛿𝑡 + 𝛼𝑥 × 𝑋𝑖𝑡 + εit

Where 𝑦𝑖𝑡 stands for total CEO pay in firm 𝑖 at time t, 𝑝𝑒𝑟𝑓𝑖𝑡 measures firm performance, 𝛾𝑖 are

independent variables that represent the firms fixed variables, 𝛿𝑡 are time fixed variables, and 𝑋𝑖𝑡

are firm characteristics and CEO’s characteristics. The coefficient 𝛽 captures the sensitivity

between performance and CEO pay.

As well, Zhou (2000) estimates the following semi-elasticity specification to examine the

relationship between pay and return and between pay and firm size.

ln (𝐶𝐸𝑂 𝑝𝑎𝑦)𝑡 = 𝑎 + 𝑏 × ln(𝑓𝑖𝑟𝑚 𝑠𝑖𝑧𝑒)𝑡 + 𝑐 × 𝑟𝑒𝑡𝑢𝑟𝑛𝑡

Chowdhury and Wang (2009) used incentive pay parentage and the natural log of

incentive pay as their dependent variables.

We extend the above models by adding a new independent variable, namely ownership

structure, and controlling for the above mentioned pay to performance relationships. Models 1-3

are intended to replicate previous research, particularly Bertrand and Mullaninathan (2001),

while models 4-6 are our main extended models. In model 1, we include the natural logarithm of

total assets (TA), ROA, M/B, D/E, and dummy year variables as our independent variables. Year

2003 serves as the base year and we add the four dummy variables Dyr04, Dyr05, Dyr06, and

Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We use the aggregated sample to

investigate the relationship between the annual bonus and firm characteristics.

𝐿𝑁(𝐶𝑖) = 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀𝐵

+ 𝛽𝐷𝐸

𝐷𝐸

+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 +𝜀𝑖,𝑡 (1)

40

In model 2, we include industry factors by adding three dummy industries to represent

the financial industry (DFin), energy industry (DEgy), and materials industry (DMat). The

resulting equation is:

𝐿𝑁(𝐶𝑖) = 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 +𝛽𝑄𝑀𝐵

+ 𝛽𝐷𝐸

𝐷𝐸

+ �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠

+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 + 𝜀𝑖,𝑡 (2)

In model 3, in order to study the impact of changes in a firm’s CEO, we use a dummy

variable (D1YC) that takes the value of 1 if there is a CEO change in the year (Transient CEO

Group) and 0 otherwise.

𝐿𝑁(𝐶𝑖) = 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷

𝐸

𝐷𝐸

+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 (3)

+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 + 𝜀𝑖,𝑡

In model 4, we include variables to control for the concentrated group (firms controlled

either by families or by institutions) and its interaction with each of the independent variables.

The introduction of concentrated ownership dummies change the results reported with model 1.

𝐿𝑁(𝐶𝑖) =

𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑀/𝐵 + 𝛽𝐷𝐸

𝐷𝐸


+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 + 𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 + 𝜔𝐷𝑐𝑜𝑛 𝐷𝑐𝑜𝑛 + 𝜔𝐷𝑐𝑜𝑛𝑅𝑂𝐴𝐷𝑐𝑜𝑛 ∗ 𝑅𝑂𝐴

+𝜔𝐷𝑐𝑜𝑛𝑚𝑏 𝐷𝑐𝑜𝑛 ∗ 𝑀/𝐵 + 𝜔𝐷𝑐𝑜𝑛𝐷/𝐸 𝐷𝑐𝑜𝑛 ∗𝐷𝐸

(4)

In model 5, we separate concentrated-controlled into two groups: family-controlled and

institution-controlled. Including widely-held, we have three variables to control for the

ownership structures. Therefore, we set two dummy variables to compare these three groups. We

compare the annual bonus in the widely-held group to that in both the family-controlled group

and the institution-controlled group, while using widely-held group as the base group.

41



𝐷𝐸


(5) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡+𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 +𝜔𝐷𝐹𝐷𝐹 + 𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 +𝜔𝐷𝐹𝑚𝑏 𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗

𝐷𝐸

+𝜔𝐷𝐼𝐷𝐼 + 𝜔𝐷𝐼𝑇𝐴 𝐷𝐼 ∗ ln(𝑇𝐴)

+𝜔𝐷𝐼𝑅𝑂𝐴𝐷𝐼 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐼𝑚𝑏𝐷𝐼 ∗ 𝑀/𝐵 + 𝜔𝐷𝐼𝐷/𝐸 𝐷𝐼 ∗𝐷𝐸

In order to investigate how annual bonus in family-controlled group is related to annual

bonus in institution-controlled group, we switch our base group, the widely-held group, to the

institution-controlled group, all else equal.



𝐷𝐸


(6) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶

+𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐹𝑚𝑏𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗𝐷𝐸

+𝜔𝐷𝑊𝑇𝐴 𝐷𝑊 ∗ ln(𝑇𝐴) +𝜔𝐷𝑊𝑅𝑂𝐴𝐷𝑊 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝑊𝑚𝑏𝐷𝑊 ∗𝑀/𝐵 + 𝜔𝐷𝑊𝐷/𝐸 𝐷𝑊 ∗𝐷𝐸

6.1 Natural logarithm of compensation as dependent variable

Table 9.1 reports the results of examining the relation between the annual bonus and

various variables. The table shows that in the absence of control for ownership structure (Models

1, 2, and 3) the annual bonus is positively related to the return on assets and to total assets. When

we divide the sample between concentrated ownership and widely held ownership additional

observations emerge. First, within the widely-held firms the bonus is positively related to total

assets and negatively related to the debt to equity ratio. Also, we find that for the widely held

firms the return on assets is not a significant determinant of the bonus. Second, the concentrated

ownership group seems to pay a higher bonus than the widely held group. Third, within the

concentrated ownership group the bonus is positively and significantly related to the return on

assets, the market to book ratio, and to the debt to equity ratio. This result suggests that the

concentrated ownership firms are linking the higher bonus payments to performance implying

support for our hypothesis that the owners of these firms are monitoring the CEOs more closely.

Fourth, within the concentrated ownership structure the relation between the bonus and asset size

42

is negative and significant which is opposite to the result we found for the widely-held firms.

Again, this result provides support to our hypothesis that the concentrated firms tie their bonuses

to more meaningful measures of performance rather than size.


Table 9.2 reports the results of examining the relation between the annual bonus after

splitting the concentrated ownership group between family-controlled and institution-controlled

firms. The table shows that both family-controlled and institution-controlled firms pay a higher

bonus than the widely held group and that the bonus is positively and significantly related to the

return on assets but negatively related to the size of assets. This result provides support to our

hypothesis that the family-controlled and institution-controlled firms tie their bonuses to more

meaningful measures of performance rather than size. Furthermore, Table 9.2 shows that the

bonus in the family-controlled firms is positively related to the debt to equity ratio but this

relation is not significant for the institution-controlled firms. Finally, Table 9.2 compares the

institution-controlled firms with the family-controlled firms. The results suggest that on average

the two groups pay similar amounts of bonuses but for the family-controlled firms the positive

relation between bonus compensation and the return on assets is stronger. This result suggests

that the level of monitoring by family-controlled firms is stronger which is consistent with our

hypothesis.


Table 10.1 reports the results of examining the relation between the incentive pay and

various variables. The table shows that in the absence of control for ownership structure (Models

1, 2, and 3) the incentive pay is either negatively related to the return on assets or the relationship

is not significant. In contrast, incentive pay seems to be positively and significantly related to

asset size and the market to book ratio. When we divide the sample between concentrated

ownership and widely held ownership (Model 4) additional observations emerge. First, within

the widely-held firms the bonus is positively related to total assets and negatively related to the

return on assets and to the debt to equity ratio. Second, the concentrated ownership group seems

to pay higher incentive compensation but this compensation is negatively related to size while its

relation to return on assets or to the market to book ratio are insignificant.


43

Table 10.2 reports the results of examining the impact of ownership structure on the

incentive pay after splitting the concentrated ownership group between family-controlled and

institution-controlled firms. The table shows that the results of Table 10.1 related to the

concentrated group can be repeated for each of the family-controlled and institution-controlled

firms taken separately. As Model 6 shows, the incentive pay offered by the family-controlled and

the institution-controlled firms does not seem to be related to performance and the levels of

incentive pay provided by the two groups seem to be similar.


Tables 10.1 and 10.2 suggest a significant industry effect in determining incentive

compensation. The energy and materials industries seem to pay significantly higher incentive

compensation than the financial industry firms. This result is consistent regardless of the

ownership structure.

6.2 Bonus and incentive compensation as percentages of total assets

When we analyze the impact of the ownership structure on the annual bonus and the

contingent compensation we find that asset size is a significant determinant of both. This result is

consistent with the findings of previous studies. In this section, we control the impact of asset

size by examining incentive compensation as a ratio of total assets.

In preparation for this analysis, we draw a graph that shows total compensation as a

function of total assets. The graph is shown in Figures 5.1-5.4. Figure 5.1 shows the graph for the

entire dataset while Figures 5.2-5.4 show the relationship for different ranges of assets. The

graphs show that for assets sizes less than $9 billion total compensation is increasing in asset size.

However, at larger asset sizes the graphs do not show a clear pattern of compensation increasing

as a function of total assets.

Figures 6.1 and 6.2 show salary plus bonus compensation as a function of total assets.

Figure 6.1 shows the graph for the entire dataset. This graph shows that bonus plus salary

increases with asset size. Figure 6.2 shows the relation for asset sizes of $20.48 million to $55

billion. For this group of firms, bonus plus salary show an increasing pattern but the rate of

increase will slow down as the size increases.

44

6.2.1. Annual bonus as a percentage of total assets

We use the ratio of annual bonus to total assets (AB-to-TA) as our dependent variable to

test the main hypotheses. Since this dependent variable is restricted at the range of [0, 1], we use

the Tobit model to mitigate the possible problem caused by censored variable. The independent

variables are similar to what included in the previous models. Table 11.1 reports the results.


The table shows that in the absence of control for ownership structure (Models 1, 2, and 3)

the annual bonus as percentage of total assets is positively related to the return on assets and

negatively related to the debt to equity ratio. When we divide the sample between concentrated

ownership and widely held ownership additional observations emerge. First, within the widely-

held firms the bonus continues to be negatively related to the debt to equity ratio but the return

on assets is no longer a significant determinant of the bonus. Second, firms within the

concentrated ownership group seem to pay a lower bonus per dollar of assets than the widely

held group. Third, within the concentrated ownership group the bonus is positively and

significantly related to the debt to equity ratio. Again, this result provides support to our

hypothesis that the concentrated firms tie their bonuses to more meaningful measures of

performance rather than size.

Table 11.2 reports the results of examining the relation between the annual bonus as a

percentage of total assets and the various independent variables after splitting the concentrated

ownership group between family-controlled and institution-controlled firms. The table shows

that both family-controlled and institution-controlled firms pay a lower bonus per dollar of assets

than the widely held group. In addition, within the family-controlled group we find that the

bonus per dollar of assets is positively and significantly related to the return on assets and to the

debt to equity ratio. Furthermore, Model 6 shows that the family-controlled firms pay lower

bonuses per dollar of assets than the institution-controlled firms and that the return on assets is a

positive and significant factor in determining the bonus. The impact of the return on assets is

stronger for the family controlled firms than for the institution controlled firms. This result

provides support to our hypothesis that the family-controlled and institution-controlled firms tie

their bonuses to more meaningful measures of performance rather than size. Furthermore, it

shows that the family-controlled firms provide more monitoring and stronger pay-performance

relation than institution-controlled firms. Finally, Table 11.2 shows that the bonus in the family-

45

controlled firms is positively related to the debt to equity ratio but this relation is negative and

significant for the institution-controlled firms.


Tables 11.1 and 11.2 suggest that new CEOs often receive lower bonus per dollar of

assets than their predecessors. This result is consistent in the various models but it is stronger

when we compare the relations across the various ownership structures. In addition, the tables

show that failing to differentiate between family-controlled and institution-controlled firms may

lead us to conclude that there is a significant industry effect suggesting that financial firms pay

lower bonus per dollar of assets. Controlling for ownership structure shows that this effect is not

significant.

6.2.2. Contingent compensation as a percentage of total assets

We use the ratio of contingent compensation to total assets (CP-to-TA) as our dependent

variable to test the main hypotheses. Since this dependent variable is restricted at the range of [0,

1], we use the Tobit model to mitigate the possible problem caused by censored variable. The

independent variables are similar to what included in the previous models. Table 12.1 reports the

results.


The table shows that in the absence of control for ownership structure (Models 1, 2, and 3)

the contingent compensation as percentage of total assets is negatively related to the return on

assets and to the debt to equity ratio but positively related to the market to book ratio. When we

divide the sample between concentrated ownership and widely held ownership these

observations continue to hold. More important, the table shows that the contingent compensation

as percentage of assets is insignificantly different from the same ratio at the widely-held firms.

Table 12.2 reports the results of examining the relation between the contingent

compensation as percentage of total assets and the various independent variables after splitting

the concentrated ownership group between family-controlled and institution-controlled firms.

The table shows that there is no evidence to suggest that the family-controlled and institution-

controlled firms pay a different percentage of contingent compensation per dollar of assets than

the widely held group.


46

Tables 12.1 and 12.2 show that there is no evidence to suggest that new CEOs receive

different contingent compensation per dollar of assets than their predecessors. This result is

consistent in all the models. In addition, the tables show that there is a strong industry effect.

When we do not control for ownership structure, Table 12.1 shows that contingent compensation

at financial institutions is lower than other industries while the energy and materials sectors pay

greater contingent compensation per dollar of assets than financial institutions or other industries.

Adding control for concentrated ownership does not change the results significantly. However,

when we differentiate between family-controlled and institution-controlled firms, Table 12.2

shows that the results change significantly. Within the family-controlled firms, the debt to equity

ratio seems to have a positive and significant impact on contingent compensation per dollar of

assets. In addition, contingent compensation at financial institutions is no longer significantly

lower than other industries. In contrast, the result related to the energy and materials sectors

continues even after dividing the concentrated ownership firms between family-controlled and

institution-controlled.

6.3 Ownership and the structure of compensation

One objective of this study is to determine whether ownership structure affects the

structure of compensation. In particular, we examine whether bonus payment and contingent

compensation as percentages of total compensation vary across the widely-held, institution-

controlled, and family-controlled firms.

6.3.1 Annual bonus as a percentage of total pay (AB-to-TP)

We use the ratio of bonus compensation to total pay (AB-to-TP) as our dependent


1], we use the Tobit model to mitigate the possible problem caused by a censored variable. The

independent variables are similar to those used in the previous models.

Table 13.1 reports the results of models 1-4. It shows that in the absence of controls for

ownership, the annual bonus as proportion of total pay is positively related to the return on assets

and negatively related to the debt to equity ratio. Table 13.2 shows that as we add controls for

ownership, the relation to the return on assets remains unchanged. Furthermore, Model 5 shows

that the ratio of bonus to total compensation in widely-held firms is not significantly different

47

from the same ratio for institution-controlled and family-controlled firms. However, the relation

between the return on assets and the ratio is strongest for the family-controlled firms while there

is no evidence to suggest that the widely-held and the institution-controlled firms display

significantly different relations.



Tables 13.1 and 13.2 also show that new CEOs receive lower proportion of their

compensation in the form of bonus payments. This result is consistent in all the models. In

addition, the tables show that there is a strong industry effect. Table 12.1 shows that when we do

not control for ownership structure, the bonus payments at financial institutions as a percentage

of total pay are higher than other industries while the energy and materials sectors pay less bonus

compensation per dollar of total compensation. As we add controls for ownership structure, the

tables show that there is no evidence to suggest that financial institutions pay significantly

different bonuses than firms in other industries. However, the result related to the energy and

materials sectors continues unchanged after we add control for ownership structure.

6.3.2 Contingent compensation as a percentage of total pay (CP-to-TP)

We use the ratio of contingent compensation to total pay (AB-to-TP) as our dependent


1], we use the Tobit model to mitigate the possible problem caused by a censored variable. The

independent variables are similar to those used in the previous models.

Table 14.1 reports the results of models 1-4. It shows that in the absence of controls for

ownership, contingent compensation as proportion of total pay is positively related to the total

assets and the market to book ratio and negatively related to the return on assets. However, the

results of Model 4 show that as we add controls for ownership concentration, the return on assets

becomes insignificant while the debt to equity ratio becomes negatively and significantly related

to the contingent compensation of the widely held firms. Table 14.2 shows that as we

differentiate between family-controlled and institution-controlled firms, we note that both pay

higher proportion of compensation as equity-based. This result is stronger for the institution-

controlled firms. Also, we observe that within the family-controlled firms the ratio of contingent

compensation to total compensation drops with the return on assets.

48



Tables 14.1 and 14.2 also show that new CEOs receive higher proportion of their

compensation in the form of contingent payments. This result is consistent in all the models. In

addition, the tables show that there is a strong industry effect. The contingent payments at

financial institutions as a percentage of total pay are lower than other industries while the energy

and materials sectors pay more contingent compensation per dollar of total compensation. This

result is consistent whether or not we control for ownership structure.

6.4 The relation between compensation and Total Market Return (TMR)

We examine whether the total market return, measured as the capital gains return on the

firm’s common shares plus the dividend yield, affects bonus compensation or stock-based

compensation. We conduct this multivariate analysis without including the market to book ratio.

Table 16 shows that the correlation between M/B and TMR is high.


The TMR data is obtained from the DataStream database. We start our analysis by using

the basic regression Model 7.11

Model 8 adds industry dummy variables to the basic model.

Model 9 adds a dummy variable to control for CEO departure and replacement, Model 10 adds a

dummy variable to compare the widely-held firms with the concentrated ownership firms. Model

11 adds dummy variables to control for ownership structure and differentiates between widely-

held, family-controlled, and institution-controlled firms. In this model, the widely-held firms

constitute the base group. Model 12 is the same as model 11 except we use the institution-

controlled firms as the base group instead of the widely held.

𝐿𝑁(𝐶𝑖) 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷𝐸

𝐷𝐸

+ ∑𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠 +𝜀𝑖,𝑡 (7)

𝐿𝑁(𝐶𝑖) 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷

𝐸

𝐷𝐸


+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 + 𝜀𝑖,𝑡

11 The dependent variable Ci is used to represent compensation and will be used to represent bonus or stock-based

compensation, TMR is total market return, and all the other variables are as defined with Equation 1.

49

𝐿𝑁(𝐶𝑖) 𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷

𝐸

𝐷𝐸


+ 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶 + 𝜀𝑖,𝑡

𝐿𝑁(𝐶𝑖)

𝛼𝑖 + 𝛽𝑇𝐴 ln(𝑇𝐴) + 𝛽𝑅𝑂𝐴𝑅𝑂𝐴 + 𝛽𝑄𝑇𝑀𝑅 + 𝛽𝐷𝐸

𝐷𝐸 + �𝛽𝑦 𝐷𝑢𝑚𝑚𝑦 𝑌𝑒𝑎𝑟𝑠

(10) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶

+𝜔𝐷𝑐𝑜𝑛 𝐷𝑐𝑜𝑛 ∗ ln(𝑇𝐴) + 𝜔𝐷𝑐𝑜𝑛𝑅𝑂𝐴𝐷𝑐𝑜𝑛 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝑐𝑜𝑛𝑚𝑏 𝐷𝑐𝑜𝑛 ∗ 𝑇𝑀𝑅 + 𝜔𝐷𝑐𝑜𝑛𝐷/𝐸 𝐷𝑐𝑜𝑛 ∗𝐷𝐸 + 𝜀𝑖,𝑡

𝐿𝑁(𝐶𝑖)



(11) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝑇𝑀𝑅𝑇𝑀𝑅 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶

+𝜔𝐷𝐹𝐷𝐹 + 𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐹𝑚𝑏 𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗𝐷𝐸

+𝜔𝐷𝐼𝐷𝐼 + 𝜔𝐷𝐼𝑇𝐴 𝐷𝐼 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐼𝑅𝑂𝐴𝐷𝐼 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐼𝑚𝑏𝐷𝐼 ∗ 𝑀/𝐵 + 𝜔𝐷𝐼𝐷/𝐸 𝐷𝐼 ∗𝐷𝐸

+𝜀𝑖,𝑡

𝐿𝑁(𝐶𝑖)



(12) + 𝛿𝐷𝐹𝑖𝑛𝐷𝐹𝑖𝑛 + 𝛿𝐷𝐸𝑔𝑦𝐷𝐸𝑔𝑦 + 𝛿𝐷𝑀𝑎𝑡𝐷𝑀𝑎𝑡 +𝛽𝑇𝑀𝑅𝑇𝑀𝑅 +𝛽𝐷1𝑌𝐶𝐷1𝑌𝐶

+𝜔𝐷𝐹𝑇𝐴 𝐷𝐹 ∗ ln(𝑇𝐴) + 𝜔𝐷𝐹𝑅𝑂𝐴𝐷𝐹 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝐹𝑚𝑏𝐷𝐹 ∗ 𝑀/𝐵 + 𝜔𝐷𝐹𝐷/𝐸 𝐷𝐹 ∗𝐷𝐸

+𝜔𝐷𝑊𝑇𝐴 𝐷𝑊 ∗ ln(𝑇𝐴) +𝜔𝐷𝑊𝑅𝑂𝐴𝐷𝑊 ∗ 𝑅𝑂𝐴 + 𝜔𝐷𝑊𝑚𝑏𝐷𝑊 ∗𝑀/𝐵 + 𝜔𝐷𝑊𝐷/𝐸 𝐷𝑊 ∗𝐷𝐸 + 𝜀𝑖,𝑡

The results of analysing the relation between bonus compensation and TMR are reported

in Tables 17.1 and 17.2. The table shows that TMR is not a significant factor in determining

bonus compensation. At the same time, the relation between bonus compensation and the other

decision and control variables remains unchanged in terms of the direction and significance. In

addition, this observation can be repeated after controlling for industry effects, CEO turnover,

and ownership structure.



50

Similarly, we examine the relation between contingent pay and TMR. The results are

reported in Tables 18.1 and 18.2. TMR does not have a significant impact on contingent pay. At

the same time, the relation between stock-based compensation and the other decision and control

variables remains unchanged in terms of the direction and significance. Adding controls for

industry effects, CEO turnover, and ownership structure does not change the insignificant

relation between stock-based compensation and total market return.



51

CHAPTER 7

Conclusions and Recommendations for Further Research

This study examines the role that ownership structure plays in the governance of large,

publicly traded firms in Canada. This subject is important given the steady rise in the levels of

CEO compensation, particularly contingent compensation, over the past two decades. We argue

that the split of CEO compensation between salary, bonus, stock-based, and other compensation

would be different across different ownership structures.

Overall, our analysis leads to many observations related to the relation between

ownership structure and compensation. First, both family-controlled and institution-controlled

firms pay higher bonus than widely-held firms and that the bonus in concentrated ownership

firms is positively and significantly related to the return on assets. This result provides support to

our hypothesis that the family-controlled and institution-controlled firms tie their bonuses to

meaningful measures of performance. Second, the analysis suggests that on average the family-

controlled and institution-controlled firms pay similar amounts of bonuses but for the family-

controlled firms the positive relation between bonus compensation and the return on assets is

stronger. This result suggests that the level of monitoring by family-controlled firms is stronger

which is consistent with our hypothesis.

We examine the relation between ownership and contingent compensation. We observe

that within the widely-held firms contingent compensation is positively related to total assets and

negatively related to the return on assets and to the debt to equity ratio. Second, the concentrated

ownership group seems to pay higher incentive compensation but this compensation is

negatively related to size while its relation to return on assets or to the market to book ratio are

insignificant. Splitting the concentrated ownership group between family-controlled and

institution-controlled firms reinforces the results reported for the concentrated ownership group.

In conclusion, the incentive pay offered by the family-controlled and the institution-controlled

firms does not seem to be related to performance and the levels of incentive pay provided by the

two groups seem to be similar.

We examine the impact of ownership structure on the compensation per dollar of assets.

Several observations emerge. First, both family-controlled and institution-controlled firms pay a

lower bonus per dollar of assets than the widely held group. Second, within the family-controlled

group we find that the bonus per dollar of assets is positively and significantly related to the

52

return on assets. Third, the family-controlled firms pay lower bonuses per dollar of assets than

the institution-controlled firms and that the return on assets is a positive and significant factor in

determining the difference in bonus. Fourth, the impact of the return on assets is stronger for the

family controlled firms than for the institution controlled firms. This result provides support to

our hypothesis that the family-controlled and institution-controlled firms tie their bonuses to

measurable metrics of performance such as return on assets. In addition, the results show that the

family-controlled firms provide more monitoring and exhibit stronger pay-performance relation

than institution-controlled firms.

Similarly, we examine the impact of ownership structure on the ratio of contingent

compensation per dollar of assets. We find that the ratio is negatively related to the return on

assets but positively related to the market to book ratio. At concentrated ownership firms,

contingent compensation as percentage of assets is insignificantly different from the same ratio

at the widely-held firms. In addition, we find no evidence to suggest that the family-controlled

and institution-controlled firms pay a different percentage of contingent compensation per dollar

of assets than the widely held group. These results suggest that stock-based compensation is not

a major tool to control the behaviour of managers but it is offered as part of a competitive

compensation package consistent with market practices.

Additional support to this conclusion is obtained when we examine the relation between

ownership structure and the ratio of bonus compensation to total compensation. The results show

that the ratio in widely-held firms is not significantly different from the same ratio at institution-

controlled and family-controlled firms. However, the relation between the return on assets and

the ratio is strongest for the family-controlled firms. In contrast, there is no evidence to suggest

that the widely-held and the institution-controlled firms display significantly different relations.

Similarly, we examine the impact of the ownership structure on the ratio of contingent

compensation to total compensation. The results show that in comparison to widely-held firms

both family-controlled and institution-controlled firms pay higher proportion of compensation as

equity-based. This result is stronger for the institution-controlled firms. Also, we observe that

within the family-controlled firms the ratio of contingent compensation to total compensation

drops with the return on assets.

53

Our results suggest a significant industry effect in determining bonus and stock-based

compensation. We find that the financial industry pays less bonus or contingent compensation

than the other industries. At the same time, the energy and materials industries seem to pay

significantly higher bonus and contingent compensation than the financial industry. When we

scale compensation by total assets, we find no differences in the bonus per dollar of assets

among the industry groups. In contrast, we find that contingent compensation per dollar of assets

at financial institutions is insignificantly lower than other industries while the energy and

materials sectors pay significantly higher contingent compensation per dollar of assets than

financial institutions or other industries.

Our analysis also compares the compensation of incoming CEOs with the compensation

of their predecessors. The results suggest that incoming CEOs often receive lower bonus per

dollar of assets than their predecessors. Consistent with these results, we also find that incoming

CEOs also receive lower proportion of their compensation in the form of bonus payments. In

contrast, the analysis shows that incoming CEOs receive higher proportion of their compensation

in the form of contingent payments.

In summary, we make significant contributions to the existing literature by comparing

bonus and stock-based compensation across widely-held, institution-controlled, and family-

controlled firms. Moreover, we compare and analyze the difference between incoming CEO

compensation and outgoing CEO compensation across the three ownership structures. Overall,

we obtain several important conclusions. First, we find that the ownership structure is very

important in determining the level and structure of CEO compensation. In particular, family-

controlled and institution-controlled firms have significantly different compensation structures.

Second, we find that bonus compensation is more associated with performance than stock-based

compensation. Third, control by a family seems to provide more effective and meaningful

monitoring of managers than control by institutions. Fourth, the structure of compensation

provided to incoming CEOs is significantly different from the structure of compensation

provided to outgoing CEOs. In particular, incoming CEOs seem to obtain less bonus

compensation and more stock-based compensation than their predecessors.

Our findings are useful for investors, academics, policy makers, and regulators. All these

stakeholders should be aware that there is room for improvement in the relationship between

managers and shareholders. Institutions do not seem to be providing active monitoring of

54

managers. Therefore, the existence of an institutional block holder does not necessarily reduce

agency costs. Control by a family seems to provide better monitoring but as documented by the

literature it has agency costs of its own. Therefore, there is need for improvement in corporate

governance practices. In addition, our results show that the compensation package is a significant

tool for influencing the CEO. However, the bonus component seems to be the only form of

compensation that is associated with performance which suggests that stock-based compensation

is not accomplishing its objectives of aligning the interests of managers and shareholders. We

speculate that two factors may be magnifying the problem. First, there is no accurate system that

sets the optimal stock-based compensation. Academics and practitioners should focus their

attention to develop such a system. Second, there is lack of information regarding the ultimate

value of stock-based compensation. The recent regulations which require the disclosure of the

value attached to stock-based compensation is a good step in this direction. We suggest that

policy makers and regulators should encourage better accounting for stock-based compensation.

This study has limitations. First, prior to 2005 Canadian firms were not required to

provide an estimate of the full monetary value of CEO stock-based compensation in their

financial statements. Thus, for some observations we use the approximate value estimated using

the Black and Scholes (1972) option pricing model. The approximations may introduce some

errors. Second, our independent variables do not include CEO characteristics, which, according

to previous studies, might have an impact on CEO compensation.

55

References

Amoako-Adu, B., Smith, B. F., 2001, Dual Class Firms: Capitalization, Ownership Structure,

and Recapitalization Back into Single Class, Journal of Banking and Finance, Vol.25, pp.

1083–1111.

Anderson, R., Reeb, D., 2003, Founding-Family Ownership and Firm Performance: Evidence

from the S&P 500, Journal of Finance, Vol.58, pp.1301–1327.

Ang, Janmes S., Cole, Rebel A., Lin, Janmes Wuh., 2002, Agency Costs and Ownership

Structure, Journal of Finance, Vol. 55, Issue 1 (Dec), pp. 81-106.

Arya, Avinash, Huey-Lian, Sun, 2004, Stock Option Repricing: Heads I Win, Tails You

Lose, Journal of Business Ethics, Vol. 50, Issue 4 (Apr), pp. 297-312.

Barenbaum, Les., and Schubert, Walt., 1993, Measuring the Value of Executive Stock Options,

Compensation & Benefits, Vol. 25 (Oct), pp. 19-24.

Baumol, W., 1959, Business Behavior, Value and Growth. New York.

Bebchuk, L. and Stole L., 1993, Do short-term objectives lead to under-or overinvestment in

contingent projects? Journal of finance, Vol. 48, pp. 719-730.

Ben-Amar, Walid. and Andre, Paul., 2006, Separation of Ownership from Control and Acquiring

Firm Performance: The Case of Family Ownership in Canada, Journal of Business

Finance & Accounting, Vol.33, No.3&4, pp.517–543.

Bennedsen, Morten et al., 2007, Inside the Family Firm: the Role of Families in Succession

Decisions and Performance, The Quarterly Journal of Economics, Vol. 122, No. 2, pp.

647-691.

Berle, Adolf A., Jr. and Gardner C, Means., 1932, The Modern Corporation and Private Property

(Macmillan, New York, NY).

Chourou, Lamia, 2010, Compensation of Owner Managers in Canadian Family-Owned

Businesses: Expropriation of Minority Shareholders, Canadian Journal of Administrative

Sciences, Vol. 27, pp. 95-106.

Chowdhury, S.D., Wang, E.Z., 2009, Institutional Activism Types and CEO Compensation: A

Time-Series Analysis of Large Canadian Corporation, Journal of Management, Vol. 35,

pp. 5-36.

56

Chrisman, James J., Chua, Jess H., and Litz, Reginald, A., 2004, Comparing the Agency Costs of

Family and Non-Family Firms: Conceptual Issues and Exploratory Evidence,

Entrepreneurship Theory and Practice, Vol. 28, Issue. 4 (May), pp. 335-354.

Ciscel, D. and Carroll T., 1980, The Determinants of Executive Salaries: An Econometric Survey,

Review of Economics and Statistics, Vol.62, No.1, pp. 7-13.

Claessens, S., Djankov, S. and Lang, L., 2000. The Separation of Ownership and Control in East

Asian Corporations, Journal of Financial Economics, Vol.58, pp. 81–112.

Daily, C. M., Johnson, J. L., Ellstrand, A. E., and Dalton, D. R., 1998, Compensation Committee

Composition as a Determinant of CEO Compensation, Academy of Management Journal,

Vol.41, pp. 209-220.

David, P., Kochhar, R., and Levitas, E., 1998, The Effect of Institutional Investors on the Level

and Mix of CEO Compensation, Academy of Management Journal, Vol.41, No.2, pp. 200.

Demsetz, H., 1983, The Structure of Ownership and the Theory of the Firm, Journal of Law and

Economic, Vol.26, pp. 375-390.

Demsetz, H., Lehn, K., 1985, The Structure of Corporate Ownership: Causes and Consequences.

Journal of Political Economy, Vol.93, pp. 1155–1177.

Faccio, Maria, and Lang, Larry P.H., 2002, The Ultimate Ownership of Western European

Corporations, Journal of Financial Economics, Vol.65, pp. 365-395

Fahlenbrach, Budiger, Stulz, Rene M., 2011, Bank CEO Incentives and the Credit Crisis, Journal

of Financial Economics, Vol.99, pp.11-26.

Fama, Eugene F., and Jensen, Michael C., 1983, Separation of Ownership and Control. Journal

of Law & Economics, Vol.26, pp. 301-325.

Fifth, Michael., Fung, Peter M.Y., Rui, Oliver M., 2006, Firm Performance, Governance

Structure, and Top Management Turnover in a Transitional Economy, Journal of

Management Studies, Vol. 43, Issue. 6 (May), pp. 1289-1330.

Florackie, C. and Ozkan, A., 2004, Agency Cost and Corporate Governance Mechanisms:

Evidence for UK firms, paper presented at European Finance Association Meeting

Gibbons, R. and Murphy, K. J., 1990, Relative Performance Evaluation for Chief Executive

Officers, Industrial and Labor Relations Review, Vol.43, No.3 (Feb), pp. 30-51.

Gray, S. R., and Cannella, A. A., 1997, The Role of Risk in Executive Compensation, Journal of

Management, Vol.23, pp. 57-540.

57

Grossman, Sanford, and Hart, Oliver., 1988, One Share-one Vote and the Market for Corporate

Control, Journal of Financial Economics, Vol.20, pp.175-202

Guidry, Flora, Leone, Andrew J., Rock, Steve, 1999, Earnings-based Bonus Plans and Earnings

Management by Business-unit Managers, Journal of Accounting and Ecomomics, Vol. 26,

pp.113-142.

Hall, Brian J., Liebman, Jeffrey B.,1998, Are CEOS Really Paid Like Bureaucrats? The

Quarterly Journal of Economics, Vol.113, Issue. 3 (Aug), pp. 653-691.

Hart, Oliver D., 1983, The Market Mechanism as an Incentive Scheme, Bell Journal of

Economics, Vol.14, pp.366-382

Hartzell, Jay C., and Starks, Laura T., 2003, Institutional Investors and Executive Compensation,

The Journal of Finance, Vol.9, No.6, pp.2351-2374.

Healy, Joanne P., and Cole, Elizabeth T., 2000, The Effect of Corporate Owneship Structure on

CEO Compensation, Journal of Deferred Compensation, Vol.3, No.2, pp. 61-72

Huson, Mark R., Parrino, Robert., and Starks, Laura T., 2002, Internal Monitoring Mechanisms

and CEO Turnover: A Long-Term Perspective, Journal of Finance, Vol. 56, Issue.6 (Dec),

pp. 2265-2297.

Jacobs, M., 1991, Short-term America: the Causes and Cures of Our Business Myopia, Harvard

Business School Press.

Jensen, Michael C., and Richard S. Ruback, 1983, The Market for Corporate Control: The

Scientific Evidence, Journal of Financial Economics, Vol.11, pp.5-50.

Jensen, Michael C., and William H. Meckling, 1976, Theory of the Firm: Managerial Behavior,

Agency Costs and Ownership Structure, Journal of Financial Economics, Vol.3, No. 4

pp.305-360.

Jensen, Michael C., and K. Murphy, 1990, Performance Pay and Top Management Incentives.

Journal of Political Economy, Vol. 98, pp. 225-264

Jeppson, Catherine T., Smith, Wayne W., Stone, Ronald S., 2009, CEO Compensation and Firm

Performance: Is There Any Relationship, Journal of Business & Economics Research, Vol.

7, No.11(Nov), pp. 81-93.

Kaplan, Steven N., 1994, Top Executive Rewards and Firm Performance: A Comparison of

Japan and the United States, The Journal of Political Economy, Vol. 102, Issue. 3 (Jun),

pp. 510-546

58

Kaplan, Steven N., and Minton, Bernadette., 2006, How Has CEO Turnover Changed?

Increasingly Performance Sensitive Boards and Increasingly Unesay CEOs, Working

paper, University of Chicago

Kato, Takao and Long, Cheryl., 2006, Executive Compensation, Firm Performance, and

Corporate Governance in China: Evidence from Firms Listed in the Shanghai and

Shenzhen Stock Exchanges, Economic Development and Cultural Change, Vol.54, No.4,

pp. 945-83.

Klein, Peter C., Daniel M. Shapiro, and Jeffrey Young, 2005, Corporate Governance, Family

Ownership and Firm Value: The Canadian Evidence, Corporate Governance: An

International Review, Vol. 13, No. 6, pp. 769-784.

Kole, Stacey R., 1997, The Complexity of Compensation Contracts, Journal of Financial

Economicas, Vol. 43, pp. 79-104.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R., 1998, Law and Finance, Journal

of Political Economy, Vol.106, pp. 1113–1155.

La Porta, R., Lopez-de-Silanes, F., Shleifer, F. and Vishny, R., 1999, Corporate Ownership

around the World, Journal of Finance, Vol.54, pp. 471–517.

Lambert, Richard A., Larcker, David F., and Verrecchia, Robert E., 1991, Portfolio

Cosiderations in Valuing Executive Compensation, Journal of Accounting Research, Vol.

29, No. 1 (Spring), pp. 129-149.

Leonard, J., 1990, Executive Pay and Firm Performance, Industrial and Labor Relations Review,

Vol.43, No.3, pp. 13-29.

Lewellen, W. and Huntsman, B. 1970, Managerial Pay and Corporate Performance, American

Economic Review, Vol.60, No.4, pp. 710-20.

Mehran, Hamid., 1995, Executive Compensation Structure, Ownership, and Firm Performance,

Journal of Financial Economics, Vol. 38, pp. 163-184.

Morck, Randall., Shleifer, Andrei., and Vishny, Robert W., 1988, Management Ownership and

Market Valuation: An Empirical Analysis, Journal of Financial Economics, Vol. 20 (Jan-

Mar), pp. 293-315.

Murphy, K. J., 1985, Corporate Performance and Managerial Remuneration: An Empirical

Analysis, Journal of Accounting and Economics, Vol.7, No.1-3, pp. 11-42.

59

Murphy, K. J., 1986, Incentives, Learning, and Compensation: a Theoretical and Empirical

Investigation of Managerial Labor Contracts, Rand Journal of Economics, Vol.17, No.1,

pp. 59-76.

Murphy, Kevin J.,1999, Executive Compensation, Hand book of Labor Economics, Vol.3, No. 2,

pp. 2485-2563.

Ozkan, Neslihan., 2007, Do Corporate Governance Mechanisms Influence CEO Compensation?

An empirical investigation of UK companies, Journal of Multinational Financial

Managerment, Vol. 17, Issue 5 (Dec.), pp. 349-364.

Roe, S. and Lee-Sing, C. R., 1996, Governance Structure, Corporate Decision-Making and Firm

Performance in North America. Working Paper, No.7, Industry Canada.

Schulze, W. S., Lubatkin, M. H., and Dino, R. N., 2003, Exploring the Agency Consequences of

Ownership Dispersion among the Directors of Private Family Firms. Academy of

Management Journal, Vol.46, No.2, pp. 179–194.

Shavell, Steven, 1979, Risk Sharing and Incentives in the Principal and Agent Relationship, The

Bell Journal of Economics, Vol.10, No.1, pp.55-73.

Shaw, Kenneth W., Zhang, May H., 2010, Is CEO Cash Compensation Punished for Poor Firm

Performance? The Accounting Review, Vol.85, Issue.3, pp. 1065-1093.

Shleifer, A and Vishny, R.W., 1986, Large Shareholders and Corporate Control. The Journal of

Political Economy, Vol.94, No.3, pp. 461-488.

Shleifer, A. and Vishny, R., 1997, A Survey of Corporate Governance, Journal of Finance, Vol

52, pp. 737-783

Singh, M. and Davidson, W. N., 2003, Agency Costs, Ownership Structure and Corporate

Governance Mechanism. Journal of Banking and Finance, Vol.27, pp. 793–816.

Smith, Michael., 1996, Shareholder Activism by Institutional Investors: Evidence from CalPERS,

Journal of Finance, Vol.51, pp. 227-252.

Tannous, George F., and Bin Cheng, (2007), Canadian Takeover Announcements and the Job

Security of Top Managers, Canadian Journal of Administrative Sciences, Volume 24,

Issue 4 (December), Pages 250-267.

Villalonga, Amit, 2006, How do Family Ownership, Control and Management Affect Firm

Value? Journal of Financial Economics, Vol.80, pp. 385-517.

60

Warner, J., R. Watts and Wruck, K., 1988, Stock Prices and Top Management Changes, Journal

of Financial Economics, Vol.20, pp. 461-92.

Zhou, Xianming., 2000, CEO Pay, Firm Size, and Corporate Performance: Evidence from

Canada, Canadian Journal of Economics, Vol.33, pp. 213–51.

61

Appendix I: Tables Table 1.1: Mean value of CEO monetary compensation in permanent CEOs firms: comparison between family controlled, institution-controlled, and widely-held firms 2003 2004 2005 2006 2007 Panel A: Family-Controlled Salary 779,140 812,010 854,452 892,431 962,269 Annual Bonus 912,965 1,700,900 1,769,673 2,330,298 2,643,697 Contingent Pay 2,758,991 3,767,955 4,406,136 4,024,199 4,477,935 Total Compensation 4,712,431 6,391,648 7,319,985 7,499,857 8,365,905 2003 2004 2005 2006 2007 Panel B: Institution-Controlled Salary 479,306 515,045 577,061 666,922 775,601 Annual Bonus 435,972 574,395 620,760 801,291 1,033,146 Contingent Pay 1,471,428 1,599,960 2,038,916 3,428,518 3,137,185 Total Compensation 2,545,116 2,888,358 3,646,281 5,031,109 5,148,305 2003 2004 2005 2006 2007 Panel C: Widely-Held Salary 487,555 530,926 552,993 580,543 607,251 Annual Bonus 599,530 694,278 869,917 1,030,685 946,963 Contingent Pay 2,128,817 2,061,454 2,242,162 2,116,396 2,133,368 Total Compensation 3,328,682 3,494,418 3,895,404 4,040,169 3,871,169 Table 1.2: Mean value of salary, annual bonus, and contingent pay as a percentage of total pay in permanent CEOs firms: comparison between family controlled, institution-controlled, and widely-held firms 2003 2004 2005 2006 2007 Panel A: Family-Controlled S-to-TP 0.1653 0.1270 0.1167 0.1190 0.1150 AB-to-TP 0.1937 0.2661 0.2418 0.3107 0.3160 CP-to-TP 0.5855 0.5895 0.6019 0.5366 0.5353 2003 2004 2005 2006 2007 Panel B: Institution-Controlled S-to-TP 0.1883 0.1783 0.1583 0.1326 0.1507 AB-to-TP 0.1713 0.1989 0.1702 0.1593 0.2007 CP-to-TP 0.5781 0.5539 0.5592 0.6815 0.6094 2003 2004 2005 2006 2007 Panel C: Widely-Held S-to-TP 0.1465 0.1519 0.1420 0.1437 0.1569 AB-to-TP 0.1801 0.1987 0.2233 0.2551 0.2446 CP-to-TP 0.6395 0.5899 0.5756 0.5238 0.5511 Notes: S-to-TP denotes salary as a percentage of total pay, AB-to-TP denotes annual bonus as a percentage of total pay, and CP-to-TP denotes contingent compensation as a percentage of total pay.

62

Table 2.1: Descriptive statistics in the transient CEO group Mean s.d. Min Med Max Mode New CEOs Salary 678,705 388,480 0 586,446 62,288,180 933,333

Annual Bonus 868,823 1,626,077 0 459,776 9,093,013 831,600

Contingent Pay 3,905,164 6,574,777 0 1,665,315 31,907,517 0

Total Pay 5,785,076 7,464,049 873,104 2,937,326 35,165,720 N/A

Old CEOs

Salary 784,580 629,311 110,500 575,000 3,450,000 600,000

Annual Bonus 652,871 993,679 0 275,252 4,816,720 0

Contingent Pay 2,459,336 5,941,128 0 637,712 31,334,307 0

Total Pay 4,194,576 6,953,677 311,886 2,329,344 37,076,147 N/A

Observations 31

Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Data sample include transient CEO firms without changing ownership structures. 3. AB-to-TP denotes annual bonus as a percentage of total pay and CP-to-TP denotes

contingent compensation as a percentage of total pay.

63

Table 2.2: Comparison of the compensation of incoming CEOs and that of their predecessors Panel A: total sample T-test: two sample assuming unequal variances DIM t-stats p-value Salary -105875 -0.797 0.215 Annual Bonus 260856 0.762 0.225 Contingent Compensation 1445827 0.908 0.184 Total 1590500 0.868 0.194 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.073 -1.692** 0.045 AB-to-TP -0.011 -0.212 0.416 CP-to-TP 0.109 1.532* 0.063 Panel B: family group T-test: two sample assuming unequal variances DIM t-stats p-value Salary -388529 -0.824 0.215 Annual Bonus 254098 0.776 0.228 Contingent Compensation 4746786 0.760 0.231 Total 4481346 0.653 0.263 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.142 -1.310 0.095 AB-to-TP 0.033 0.350 0.364 CP-to-TP 0.210 1.074 0.141 Panel C: Institution group T-test: two sample assuming unequal variances DIM t-stats p-value Salary -17245 0.151 0.441 Annual Bonus 153785 0.425 0.338 Contingent Compensation 166215 0.738 0.235 Total 1615033 1.248 0.116 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.048 -0.766 0.222 AB-to-TP -0.037 -0.402 0.344 CP-to-TP 0.131 1.524* 0.064 Panel D: Widely-held group T-test: two sample assuming unequal variances DIM t-stats p-value Salary -30748 -0.248 0.404 Annual Bonus -49925 -0.181 0.30 Contingent Compensation -456202 -0.323 0.375 Total -278123 -0.145 0.443 Z-test: two sample for means DIM z-stats p-value S-to-TP -0.058 -0.792 0.214 AB-to-TP -0.008 -0.108 0.457 CP-to-TP 0.002 0.145 0.442 Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Data sample include transient CEO firms without changing ownership structures. 3. DIM denotes differences in mean.

64

Table 3: Descriptive statistic related to the dependent and control variables Mean s.d. Min Med Max Mode Dependent Variables Annual Bonus 913,019 1,465,376 0 500,000 12,929,728 0

Contingent Pay 2,885,466 5,149,953 0 1,206,930 63,815,034 0

Control Variables

ln(TA) 21.77 1.74 16.83 21.56 27.12 21.61

ROA 4.17 7.13 -44.97 3.70 32.02 1.63

M/B 2.73 1.71 0.55 2.35 16.49 1.89

D/E (%) 85.77 135.43 0 44.93 1383.31 0

Notes: 1. Ln (TA) denotes capital expenditure, ROA denotes return on assets, M/B denotes

market-to-book value, and D/E denotes debt to equity ratio. 2. Data include both permanent CEO firms and transient CEO firms

65

Table 4: Descriptive statistic related to the dependent and control Variables in three different ownership structures

Panel A: Family-controlled group Dependent Variables Mean s.d. Min Med Max Mode Annual Bonus 1,460,940 2,431,252 0 748,750 12,929,728 0 Contingent pay 5,489,968 9,914,024 0 1,326,253 63,815,034 0 Control Variables Ln(TA) 22.76 1.41 20.47 22.59 25.61 N/A ROA 2.54 4.61 -17.29 1.90 12.27 N/A M/B 2.29 1.21 0.55 1.89 7.68 1.89 D/E (%) 139.76 241.91 0 53.45 1383.31 0 Panel B: Institution-controlled group Dependent Variables Mean s.d. Min Med Max Mode Annual Bonus 824,449 1,317,417 0 475,000 12,807,300 0 Contingent pay 2,423,975 3,312,135 0 1,287,596 27,404,855 0 Control Variables ln(TA) 21.41 1.37 16.83 21.41 24.73 N/A ROA 4.93 7.91 -41.91 4.18 32.02 N/A M/B 2.89 1.84 0.61 2.36 12.43 1.51 D/E (%) 65.66 72.80 0 47.17 437.66 0 Panel C: Widely-Held Group Dependent Variables Mean s.d. Min Med Max Mode Annual Bonus 775,703 944,806 0 476,937 5,000,000 0 Contingent pay 2,271,854 3,066,349 0 1,087,309 17,832,455 0 Control Variables ln(TA) 21.65 2.01 17.65 21.24 27.12 21.61 ROA 4.10 7.07 -44.97 3.95 23.48 -0.31 M/B 2.76 1.73 0.60 2.40 16.49 1.60 D/E (%) 84.21 115.01 0 40.99 991.55 0 Notes: 1. Ln (TA) denotes capital expenditure, ROA denotes return on assets, M/B denotes

market-to-book ratio, and D/E denotes debt to equity ratio. 2. Data include both permanent CEO firms and transient CEO firms. Observations are 96

in family-controlled group, 228 in institution-controlled group, and 236 in widely held group.

66

Table 5: Descriptive statistic related to the dependent and control variables in different industries Panel A: Financial Industry Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 1,554,841 1,484,468 0 1,400,000 5,000,000 0 Contingent pay 4,402,151 5,583,377 0 1,935,630 20,479,090 0 Control Variables Ln(TA) 23.89 2.16 18.87 23.69 27.12 N/A ROA 2.34 2.73 -1.98 1.41 12.08 1.02 M/B 2.50 1.13 0.76 2.42 6.98 0.76 D/E (%) 114.55 134.36 0 47.68 480.70 0 Panel B: Energy Industry Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 371,484 488,355 0 287,500 3,874,900 0 Contingent pay 1,966,984 2,468,270 0 883,167 11,711,567 0 Control Variables ln(TA) 20.91 1.33 17.65 20.85 23.62 N/A ROA 6.31 7.26 -11.17 5.39 32.02 N/A M/B 2.66 1.25 0.64 2.36 6.39 N/A D/E (%) 89.81 92.15 0 59.33 437.66 0 Panel C: Material Industry Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 690,189 770,581 0 513,500 4,822,429 0 Contingent pay 2,446,107 2,653,612 0 1,556,225 10,382,308 0 Control Variables ln(TA) 21.21 1.17 18.42 21.23 23.94 N/A ROA 5.97 7.11 -12.72 5.04 30.74 N/A M/B 3.06 1.90 0.60 2.49 9.34 3.19 D/E (%) 40.40 50.15 0 24.80 300.16 0 Panel D: Other Industries Dependent Variables Mean s.d. Min Med Max Mode Bonus pay 970,372 1,775,991 0 548,000 12,929,728 0 Contingent pay 2,865,062 6,228,979 0 1,168,075 63,815,034 0 Control Variables ln(TA) 21.57 1.29 16.83 21.53 25.51 N/A ROA 3.34 7.77 -44.97 3.98 24.10 1.63 M/B 2.70 1.90 0.55 2.24 16.49 2.22 D/E (%) 93.32 164.71 0 47.04 1383.31 0 Notes: 1. Ln (TA) denotes capital expenditure, ROA denotes return on assets, M/B denotes

market-to-book value, and D/E denotes debt to equity ratio. 2. Data include both permanent CEO firms and transient CEO firms. Observations are 90

in the financial industry, 90 in the energy industry, 110 in the material industry, and 270 in the other industries.

67

Table 6.1: T-test: two sample assuming unequal variances (monetary terms across three ownerships) Family/Institution Family/Widely Held Institution/Widely Held Annual Bonus 2.420** 2.680*** 0.457 Contingent Pay 2.961*** 3.120*** 0.513 Table 6.2: Descriptive statistics mean value annual bonus and contingent pay Panel A: Mean Family Institution Widely Held AB-to-TP 0.236 0.216 0.219 CP-to-TP 0.456 0.468 0.429 Panel B: z-test: Family/Institution Family/Widely Held Institution/Widely Held AB-to-TP 0.789 0.680 -0.158 CP-to-TP -0.334 0.712 1.442 Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Z-test: two Sample for Means (different ownership structures) 3. Data include both permanent CEO firms and transient CEO firms 4. AB-to-TP denotes annual bonus as a percentage of total pay and CP-to-TP denotes

contingent compensation as a percentage of total pay.

68

Table 7: Descriptive statistics mean value for the control variables Family/Institution Family/Widely Held Institution/Widely Held Ln(TA) 7.595*** 5.679*** -1.172 ROA -3.406*** -2.38*** 1.191 M/B -3.449*** -2.819*** 0.770 D/E 2.907*** 2.153** 2.200** Note: 1. ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. 2. Z-test: two Sample for Means (different ownership structures) 3. Data include both permanent CEO firms and transient CEO firms

69

Table 8.1: Correlation of variables in the aggressive sample Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1 ln(CP) 0.2139 1 ln(TA) 0.1428 0.2056 1 ROA 0.2371 -0.0419 -0.0642 1 M/B 0.0547 0.0629 -0.0919 0.0944 1 D/E -0.0687 0.0563 0.3141 -0.1600 0.0120 1 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1 CP-to-TP -0.5885 1 ln(TA) -0.0167 0.1723 1 ROA 0.1830 -0.0651 -0.0642 1 M/B -0.0312 0.0961 -0.0919 0.0944 1 D/E -0.1240 0.0353 0.3141 -0.1560 0.0120 1 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1 CP-to-TA 0.1991 1 ROA 0.1879 -0.0778 1 M/B 0.0278 0.2189 0.0944 1 D/E -0.1863 -0.0948 -0.1600 0.0120 1 Notes: 1. data include both permanent CEOs firms and transient CEO firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent

pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(ta) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.

70

Table 8.2: Correlation of variables in the family-controlled firms Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1.0000

ln(CP) 0.0362 1.0000 ln(TA) -0.0612 0.1276 1.0000

ROA 0.4371 -0.2052 -0.2257 1.0000 M/B 0.1011 -0.0217 0.0811 0.2520 1.0000

D/E -0.1287 0.0305 -0.0428 -0.2187 0.3237 1.0000 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1.0000

CP-to-TP -0.6758 1.0000 ln(TA) -0.1354 0.1640 1.0000

ROA 0.3638 -0.3199 -0.2257 1.0000 M/B -0.0413 0.0477 0.0811 0.2520 1.0000

D/E -0.2121 0.0104 -0.0428 -0.2187 0.3237 1.0000 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1.0000

CP-to-TA 0.0885 1.0000 ROA 0.2971 -0.2699 1.0000

M/B -0.0725 0.0061 0.2520 1.0000 D/E -0.0360 0.1140 -0.2187 0.3237 1.0000

Notes: 1. data include family controlled firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.

71

Table 8.3: Correlation of variables in the institution-controlled firms Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1.0000

ln(CP) 0.1905 1.0000 ln(TA) -0.0238 0.0098 1.0000

ROA 0.3605 0.0363 0.0872 1.0000 M/B 0.1107 0.1095 -0.1661 0.0600 1.0000

D/E -0.1242 0.0088 0.2027 -0.0283 0.0926 1.0000 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1.0000

CP-to-TP -0.6347 1.0000 ln(TA) 0.0197 0.0673 1.0000

ROA 0.1778 -0.0066 0.0872 1.0000 M/B -0.0187 0.1341 -0.1661 0.0600 1.0000

D/E -0.1963 0.0904 0.2027 -0.0283 0.0926 1.0000 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1.0000

CP-to-TA 0.1577 1.0000 ROA 0.2560 -0.1152 1.0000

M/B 0.0418 0.2539 0.0600 1.0000 D/E -0.2564 0.0505 -0.0283 0.0926 1.0000

Notes: 1. data include institution controlled firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.

72

Table 8.4: Correlation of variables in the widely-held firms Panel A ln (AB) ln (CP) ln (TA) ROA M/B D/E ln(AB) 1.0000

ln(CP) 0.3095 1.0000 ln(TA) 0.3199 0.3782 1.0000

ROA 0.0750 -0.0847 -0.0933 1.0000 M/B -0.0106 0.0424 -0.0211 0.0750 1.0000

D/E 0.0066 0.1376 0.5780 -0.2474 -0.1536 1.0000 Panel B AB-to-TP CP-to-TP ln(TA) ROA M/B D/E AB-to-TP 1.0000

CP-to-TP -0.5078 1.0000 ln(TA) -0.0250 0.2611 1.0000

ROA 0.1581 -0.0645 -0.0933 1.0000 M/B -0.0321 0.0758 -0.0211 0.0750 1.0000

D/E -0.0528 0.0419 0.5780 -0.2474 -0.1536 1.0000 Panel C AB-to-TA CP-to-TA ROA M/B D/E AB-to-TA 1.0000

CP-to-TA 0.2269 1.0000 ROA 0.1233 -0.0635 1.0000

M/B 0.0054 0.1709 0.0750 1.0000 D/E -0.2504 -0.2474 -0.2474 -0.1536 1.0000

Notes: 1. data include widely-held firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, AB-to-TP denotes annual bonus as a percentage of total pay, CP-to-TP denotes contingent pay as a percentage of total pay, AB-to-TA denotes annual bonus as a percentage of total assets, CP-to-TA denotes contingent pay as a percentage of total assets, TP-to-TA denotes total pay as a percentage of total assets, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market-to-book ratio, D/E denotes debt to equity ratio.

73

Table 9.1: The impact of ownership structure on the natural log of annual bonus (OLS) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α -1.6501 -0.5591 -3.3069 -1.0219 -3.3017 -1.0202 -18.5654 -4.4370*** Ln(TA) 0.5147 3.7519*** 0.5964 4.0201*** 0.6009 4.0652*** 1.4043 7.2911*** ROA 0.1576 5.3582*** 0.1538 5.0025*** 0.1505 4.9438*** 0.0260 0.5016 M/B 0.1318 1.0031 0.1212 0.9246 0.1078 0.8259 -0.2208 -0.9994 D/E -0.0033 -1.6370 -0.0032 -1.5868 -0.0032 -1.5858 -0.0122 -4.2756*** Dyr04 0.7641 1.1134 0.7476 1.0848 0.7997 1.1543 0.9134 1.3562 Dyr05 1.2122 1.8491* 1.1897 1.8070* 1.3130 1.9933** 1.6056 2.5129** Dyr06 1.2876 1.9792** 1.2503 1.8996* 1.2995 1.9691** 1.4806 2.3589** Dyr07 1.0417 1.5481 0.9905 1.4625 0.9804 1.4440 1.3436 2.0263** Dfin -0.6837 -0.9580 -0.7181 -1.0061 -1.5933 -2.0838** Degy -0.2428 -0.4346 -0.2888 -0.5144 -0.6361 -1.1167 Dmat 0.4733 0.9613 0.4779 0.9708 -0.7076 -1.3563 D1yc -1.2112 -1.3262 -1.2610 -1.4504 Dcon 28.9372 4.4918*** Dcon*ln(TA) -1.4380 -4.8760*** Dcon*ROA 0.2053 3.2728*** Dcon*M/B 0.4452 1.6966* Dcon*D/E 0.0088 2.4625** Adjusted R-square 0.087 0.086 0.088 0.150 F-statistic 7.614 5.762 5.481 6.822 Observations 560 560 560 560

74

Table 9.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 T-stat 6 T-stat α -19.7558 -4.6390*** 11.9361 1.9702** Ln(TA) 1.4552 7.4315*** -0.0972 -0.3482 ROA 0.0241 0.4649 0.2009 5.4715*** M/B -0.2094 -0.9423 0.2223 1.5724 D/E -0.0121 -4.2471*** -0.0088 -1.8478* Dyr04 0.9867 1.4704 0.9867 1.4704 Dyr05 1.5775 2.4698** 1.5775 2.4698** Dyr06 1.4743 2.3639** 1.4743 2.3639** Dyr07 1.3295 2.0242** 1.3295 2.0242** Dfin -1.8389 -2.3194** -1.8389 -2.3194** Degy -0.2803 -0.4880 -0.2803 -0.4880 Dmat -0.5409 -0.9991 -0.5409 -0.9991 D1yc -1.2526 -1.4444 -1.2526 -1.4444 DF 22.3362 2.3589** -9.3557 -0.8522 DF*ln(TA) -1.1513 -2.6934** 0.4011 0.8051 DF*ROA 0.4937 4.2209*** 0.3169 2.8597*** DF*M/B 0.2789 0.4921 -0.1528 -0.2790 DF*D/E 0.0105 2.6905*** 0.0072 1.3207 DI 31.6919 4.0301*** DI*ln(TA) -1.5524 -4.2574*** DI*ROA 0.1768 2.7932*** DI*M/B 0.4316 1.6151 DI*D/E 0.0033 0.6113 Adjusted R-square 0.159 0.159 F-statistic 5.799 5.799 Observations 560 560

75

Table 10.1: The impact of ownership structure on the natural log of contingent compensation (OLS) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α -4.3253 -1.6114 -13.568 -3.8456*** -13.5724 -3.8310*** -26.7482 -6.0524*** Ln(TA) 0.6792 5.4893*** 1.1033 6.7929*** 1.0995 6.7473*** 1.7452 8.5551*** ROA -0.0315 -1.0019 -0.0656 -1.9987** -0.0629 -1.9135* -0.0859 -1.9027* M/B 0.2697 1.9932** 0.2556 1.8801* 0.2667 1.9627* 0.1444 0.5971 D/E -0.0007 -0.3462 -0.0011 -0.5287 -0.0011 -0.5303 -0.0059 -1.9502* Dyr04 0.8207 1.0619 0.7383 0.9760 0.6948 0.9187 0.7988 1.0679 Dyr05 1.3577 1.8249* 1.2298 1.7011* 1.1270 1.5434 1.2912 1.7714* Dyr06 0.6171 0.7877 0.4328 0.5684 0.3917 0.5142 0.5374 0.7057 Dyr07 0.9187 1.1794 0.6707 0.8690 0.6791 0.8780 0.9242 1.2006 Dfin -2.2535 -2.6250*** -2.2248 -2.5719** -2.9027 -3.2105*** Degy 1.8168 2.8715*** 1.8551 2.9257*** 1.7478 2.7077** Dmat 2.1719 3.8515*** 2.1681 3.8380*** 1.3828 2.3468** D1yc 1.0094 1.2034 0.9620 1.1756 Dcon 25.8375 4.3670*** Dcon*ln(TA) -1.2128 -4.4807*** Dcon*ROA 0.0359 0.5624 Dcon*M/B 0.1252 0.4352 Dcon*D/E 0.0043 1.0701 Adjusted R-squared 0.043 0.088 0.088 0.108 F-statistic 4.152 5.927 5.539 4.995 Observations 560 560 560 560

76

Table 10.2: Natural log of contingent compensation in widely-held, institution-controlled, and family-controlled firms (OLS) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 T-stat 6 T-stat α -26.9059 -6.0247*** 3.0363 0.5229 Ln(TA) 1.7523 8.5198*** 0.3571 1.3103 ROA -0.0871 -1.9162* -0.0224 -0.4271 M/B 0.1386 0.5689 0.2433 1.4815 D/E -0.0061 -1.9752** -0.0044 -1.0231 Dyr04 0.7966 1.0531 0.7966 1.0531 Dyr05 1.3935 1.9001* 1.3935 1.9001 Dyr06 0.6226 0.8141 0.6226 0.8141 Dyr07 1.0040 1.2973 1.0040 1.2973 Dfin -2.9082 -3.2275*** -2.9082 -3.2275*** Degy 1.8000 2.6277*** 1.8000 2.6277*** Dmat 1.2242 2.0911** 1.2242 2.0911** D1yc 0.8280 1.0160 0.8280 1.0160 DF 18.7768 2.1508** -11.1654 -1.1608 DF*ln(TA) -0.9050 -2.3270** 0.4902 1.1330 DF*ROA -0.1455 -1.1179 -0.2102 -1.5696 DF*M/B 0.2823 0.5404 0.1776 0.3602 DF*D/E 0.0044 0.8960 0.0028 0.4853 DI 29.9422 4.2630*** DI*ln(TA) -1.3951 -4.2550*** DI*ROA 0.0647 0.9386 DI*M/B 0.1046 0.3532 DI*D/E 0.0016 0.2972 Adjusted R-squared 0.108 0.108 F-statistic 4.076 4.076 Observations 560 560

77

Table 11.1: The impact of ownership structure on the annual bonus as a proportion of total assets (Tobit) In model1, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α 0.0002 2.5188** 0.0002 3.0105*** 0.0002 3.1780*** 0.0005 4.1727*** ROA 0.00002 3.6438*** 0.00002 3.4572*** 0.00002 3.3909*** 0.000008 1.0858 M/B 0.000009 0.5368 0.000006 0.3944 0.000005 0.2950 -0.00003 -0.9553 D/E -0.000001 -2.9567*** -0.000001 -2.8902*** -0.000001 -2.9014*** -0.000002 -4.7242*** Dyr04 0.0001 1.2307 0.0001 1.2168 0.0001 1.2836 0.00009 1.1885 Dyr05 0.0001 1.7587* 0.0001 1.7671* 0.0002 1.9174* 0.0002 2.0230** Dyr06 0.00008 1.1121 0.00008 1.1129 0.0001 1.1806 0.0001 1.2063 Dyr07 0.0001 1.6496* 0.0001 1.6536* 0.0001 1.6172 0.0001 1.7764* Dfin -0.0002 -2.6492*** -0.0002 -2.6927** -0.0001 -1.8260* Degy -0.0001 -0.8092 -0.00007 -0.8914 -0.0001 -1.4529 Dmat -0.000004 -0.0658 -0.000004 -0.0620 -0.000002 -0.0238 D1yc -0.0002 -1.9539* -0.0002 -1.9446* Dcon -0.0004 -3.0213*** Dcon*ROA 0.00002 1.6394 Dcon*M/B 0.00005 1.4442 Dcon*D/E 0.000001 2.6616*** Adjusted R-square 0.052 0.056 0.058 0.075 Akaike info criterion -9.802 -9.802 -9.803 -9.813 Schwarz criterion -9.733 -9.709 -9.702 -9.682 Observations 560 560 560 560

78

Table 11.2: Annual bonus as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α 0.0005 4.1301*** 0.0002 2.4275** ROA 0.000008 1.0944 0.00002 2.8867*** M/B -0.00003 -0.9758 0.00002 1.2283 D/E -0.000002 -4.8765*** -0.000002 -3.6236*** Dyr04 0.0001 1.2943 0.0001 1.2943 Dyr05 0.0002 2.1752** 0.0002 2.1752** Dyr06 0.0001 1.3939 0.0001 1.3939 Dyr07 0.0001 1.9003* 0.0001 1.9003* Dfin -0.00009 -1.2541 -0.00009 -1.2541 Degy -0.00008 -1.0340 -0.00008 -1.0340 Dmat -0.00002 -0.2825 -0.00002 -0.2825 D1yc -0.0002 -2.3242** -0.0002 -2.3242** DF -0.0005 -3.1721*** -0.0002 -1.6692* DF*ROA 0.00004 2.7022*** 0.00003 1.9143* DF*M/B -0.00003 -0.6156 -0.00008 -1.8212* DF*D/E 0.000002 4.2951*** 0.000002 3.5371*** DI -0.0003 -1.9404* DI*ROA 0.00001 1.2409 DI*M/B 0.00005 1.4602 DI*D/E 0.0000001 -0.1166 Adjusted R-square 0.085 0.085 Akaike info criterion -9.822 -9.822 Schwarz criterion -9.660 -9.660 Observations 560 560

79

Table 12.1: The impact of ownership structure on the contingent compensation as a proportion of total assets (Tobit) In model1, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 Z-stat 2 Z-stat 3 Z-stat 4 Z-stat α 0.0006 2.0627** 0.0004 1.5622 0.0004 1.3873 0.0006 1.7319* ROA -0.00004 -2.2782** -0.00005 -3.0313*** -0.00005 -2.9291*** -0.00005 -1.8982* M/B 0.0003 4.1542*** 0.0003 4.2522*** 0.0003 4.3373*** 0.0002 2.3265** D/E 0.0000 -3.1833*** -0.000001 -2.2295** 0.00000 -2.2581** 0.0000 -2.5873** Dyr04 -0.0002 -0.4758 -0.0002 -0.4857 -0.0002 -0.5739 -0.0002 -0.5519 Dyr05 -0.0001 -0.3165 -0.0001 -0.2941 -0.0002 -0.4981 -0.0002 -0.4436 Dyr06 -0.0006 -1.7066* -0.0005 -1.6866* -0.0006 -1.7607* -0.0006 -1.7169* Dyr07 -0.0006 -1.8566* -0.0006 -1.8565* -0.0006 -1.8633* -0.0005 -1.7501* Dfin -0.0006 -2.6783*** -0.0006 -2.5737** -0.0005 -1.9563* Degy 0.0008 2.5131** 0.0008 2.6140*** 0.0008 2.5224** Dmat 0.0008 2.9579*** 0.0008 2.9890*** 0.0008 2.8227*** D1yc 0.0007 1.5803 0.0007 1.5993 Dcon -0.0004 -0.9990 Dcon*ROA -0.000003 -0.1018 Dcon*M/B 0.0001 0.6395 Dcon*D/E 0.000002 1.2832 Adjusted R-square 0.051 0.088 0.095 0.093 Akaike info criterion -7.416 -7.454 -7.458 -7.446 Schwarz criterion -7.347 -7.361 -7.357 -7.315 Observations 560 560 560 560

80

Table 12.2: Contingent compensation as a proportion of total assets in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α 0.0007 1.8813* 0.0004 1.0006 ROA 0.0000 -1.8756* -0.00005 -2.1724** M/B 0.0002 2.2660** 0.0003 3.4936*** D/E -0.000003 -2.8055*** 0.0000002 0.0765 Dyr04 -0.0002 -0.5644 -0.0002 -0.5644 Dyr05 -0.0002 -0.4225 -0.0002 -0.4225 Dyr06 -0.0005 -1.6851* -0.0005 -1.6851* Dyr07 -0.0005 -1.6869* -0.0005 -1.6869* Dfin -0.0004 -1.5971 -0.0004 -1.5971 Degy 0.0007 2.0223** 0.0007 2.0223** Dmat 0.0007 2.4898** 0.0007 2.4898** D1yc 0.0007 1.5850 0.0007 1.5850 DF -0.0004 -1.2073 -0.0002 -0.4803 DF*ROA 0.0000 -0.4345 -0.00001 -0.3104 DF*M/B -0.0001 -0.8188 -0.0002 -1.5568 DF*D/E 0.000003 2.1373** -0.0000003 -0.0961 DI -0.0003 -0.5942 DI*ROA -0.000004 -0.1323 DI*M/B 0.0001 0.6431 DI*D/E 0.000003 0.8801 Adjusted R-square 0.095 0.095 Akaike info criterion -7.443 -7.443 Schwarz criterion -7.281 -7.281 Observations 560 560

81

Table 13.1: The impact of ownership structure on the annual bonus as a proportion of total pay (Tobit) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% level. Independent Variable Model 1 T-stat 2 T-stat 3 T-stat 4 T-stat α 0.0905 0.7460 0.3643 2.3812 ** 0.3700 2.4166** 0.2155 1.0415 Ln(TA) 0.0032 0.5911 -0.0093 -1.3424 -0.0092 -1.3249 -0.0007 -0.0684 ROA 0.0063 4.7976*** 0.0074 5.4197 *** 0.0071 5.2828 *** 0.0049 2.1817 ** M/B -0.0045 -0.7842 -0.0047 -0.8131 -0.0057 -0.9953 -0.0129 -1.4743 D/E -0.0002 -2.0458 ** -0.0002 -1.9007 * -0.0002 -1.9165 * -0.0002 -1.2373 Dyr04 0.0240 0.8049 0.0263 0.8986 0.0298 1.0196 0.0324 1.0907 Dyr05 0.0432 1.4574 0.0476 1.6387 0.0567 1.9388 * 0.0622 2.0955 ** Dyr06 0.0760 2.4716 ** 0.0819 2.7195 *** 0.0852 2.8249 *** 0.0875 2.8999 *** Dyr07 0.0624 2.0556 ** 0.0698 2.3438 ** 0.0685 2.3052 ** 0.0735 2.4436 ** Dfin 0.0636 1.7578 * 0.0613 1.6849 * 0.0463 1.1563 Degy -0.0712 -2.7102 *** -0.0750 -2.8698 *** -0.0811 -3.0798 *** Dmat -0.0425 -1.9338 * -0.0423 -1.9413 * -0.0625 -2.6141 *** D1yc -0.0941 -2.8782 *** -0.0965 -2.9541 *** Dcon 0.3180 1.1278 Dcon*ln(TA) -0.0163 -1.2575 Dcon*ROA 0.0037 1.3641 Dcon*M/B 0.0118 1.0327 Dcon*D/E 0.0000 0.1033 Adjusted R-square 0.045 0.075 0.083 0.076 Akaike info criterion 0.159 0.143 0.135 0.143 Schwarz criterion 0.236 0.244 0.243 0.299 Observations 560 560 560 560

82

Table 13.2: Annual bonus as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α 0.1923 0.9250 0.4096 1.4503 Ln(TA) 0.0002 0.0243 -0.0113 -0.8612 ROA 0.0048 2.1770 ** 0.0072 4.2559 *** M/B -0.0124 -1.4140 0.0032 0.3779 D/E -0.0002 -1.2015 -0.0005 -2.2467 ** Dyr04 0.0339 1.1379 0.0339 1.1379 Dyr05 0.0599 2.0119 ** 0.0599 2.0119 ** Dyr06 0.0857 2.8717 *** 0.0857 2.8717 *** Dyr07 0.0710 2.3805 ** 0.0710 2.3805 ** Dfin 0.0435 1.0688 0.0435 1.0688 Degy -0.0671 -2.5134 ** -0.0671 -2.5134 ** Dmat -0.0536 -2.2007 ** -0.0536 -2.2007 ** D1yc -0.0986 -2.9901 *** -0.0986 -2.9901 *** DF 0.3970 0.9982 0.1797 0.3883 DF*ln(TA) -0.0184 -1.0465 -0.0069 -0.3359 DF*ROA 0.0178 2.9426 *** 0.0155 2.6423 *** DF*M/B -0.0134 -0.6925 -0.0290 -1.4962 DF*D/E 0.0001 0.6584 0.0004 1.6909 DI 0.2173 0.6146 DI*ln(TA) -0.0115 -0.6970 DI*ROA 0.0023 0.8452 DI*M/B 0.0156 1.2706 DI*D/E -0.0003 -1.0204 Adjusted R-square 0.083 0.083 Akaike info criterion 0.140 0.140 Schwarz criterion 0.326 0.326 Observations 560 0.083

83

Table 14.1: The impact of ownership structure on the contingent compensation as a proportion of total pay (Tobit) In model1, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 2, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 3, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 4, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 1 Z-stat 2 Z-stat 3 Z-stat 4 Z-stat α -0.4562 -2.6005*** -1.1850 -5.4820*** -1.1814 -5.4880*** -1.6976 -5.9681*** Ln(TA) 0.0381 4.8804*** 0.0711 7.2053*** 0.0702 7.1432*** 0.0953 7.3733*** ROA -0.0032 -1.5541 -0.0062 -2.9890*** -0.0056 -2.7216*** -0.0052 -1.4761 M/B 0.0233 2.6201*** 0.0233 2.6037*** 0.0254 2.8972*** 0.0200 1.3787 D/E -0.0001 -0.7813 -0.0001 -1.1694 -0.0001 -1.2016 -0.0004 -2.5275** Dyr04 0.0254 0.5131 0.0188 0.3917 0.0100 0.2115 0.0144 0.3069 Dyr05 0.0389 0.8017 0.0277 0.5984 0.0066 0.1437 0.0129 0.2775 Dyr06 -0.0258 -0.5159 -0.0406 -0.8612 -0.0490 -1.0503 -0.0427 -0.9081 Dyr07 -0.0173 -0.3562 -0.0364 -0.7680 -0.0349 -0.7399 -0.0242 -0.5094 Dfin -0.1570 -3.2061*** -0.1501 -3.0390*** -0.1686 -3.1904*** Degy 0.1953 4.5852*** 0.2026 4.8187*** 0.2011 4.7260*** Dmat 0.1546 4.1607*** 0.1538 4.1914*** 0.1263 3.2045*** D1yc 0.1996 3.5715*** 0.1976 3.5261*** Dcon 0.9688 2.5093** Dcon*ln(TA) -0.0452 -2.5691** Dcon*ROA -0.0009 -0.2169 Dcon*M/B 0.0058 0.3149 Dcon*D/E 0.0003 1.4012 Adjusted R-square 0.049 0.109 0.134 0.132 Akaike info criterion 1.035 0.962 0.942 0.949 Schwarz criterion 1.112 1.062 1.051 1.096 Observations 560 560 560 560

84

Table 14.2: Contingent compensation as a proportion of total pay in widely-held, institution-controlled, and family-controlled firms (Tobit) In model 5, use total assets (ln (TA)), return on assets (ROA), market to book ratio (M/B), debt to equity ratio (D/E), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). We add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family-controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 5 Z-stat 6 Z-stat α -1.6953 -5.9276*** -0.5412 -1.4740 Ln(TA) 0.0953 7.3434*** 0.0419 2.4206** ROA -0.0053 -1.4947 -0.0040 -1.3962 M/B 0.0195 1.3447 0.0205 1.6673* D/E -0.0004 -2.5667** -0.0001 -0.4769 Dyr04 0.0112 0.2363 0.0112 0.2363 Dyr05 0.0183 0.3939 0.0183 0.3939 Dyr06 -0.0394 -0.8419 -0.0394 -0.8419 Dyr07 -0.0226 -0.4750 -0.0226 -0.4750 Dfin -0.1699 -3.1934*** -0.1699 -3.1934*** Degy 0.1978 4.4247*** 0.1978 4.4247*** Dmat 0.1176 2.9843*** 0.1176 2.9843*** D1yc 0.1908 3.4065*** 0.1908 3.4065*** DF 0.9299 1.6910* -0.2242 -0.3649 DF*ln(TA) -0.0454 -1.8756* 0.0079 0.2889 DF*ROA -0.0197 -2.4735** -0.0209 -2.6942*** DF*M/B 0.0480 1.5943 0.0470 1.5943 DF*D/E 0.0002 0.7554 -0.0001 -0.2872 DI 1.1541 2.4893** DI*ln(TA) -0.0534 -2.4692** DI*ROA 0.0013 0.2805 DI*M/B 0.0010 0.0519 DI*D/E 0.0003 0.8486 Adjusted R-square 0.136 0.136 Akaike info criterion 0.954 0.954 Schwarz criterion 1.140 1.140 Hannan-Quinn criter. 1.027 1.027 Observations 560 560

85

Table 15: Correlation of variables in the family-controlled, institution-controlled, and widely-held firms Panel A ln (AB) ln (CP) ln (ta) ROA M/B D/E TMR ln(AB) 1 ln(CP) 0.2139 1 ln(ta) 0.1428 0.2056 1 ROA 0.2371 -0.0419 -0.0642 1 M/B 0.0547 0.0629 -0.0919 0.0944 1 D/E -0.0687 0.0563 0.3141 -0.1600 0.0120 1 TMR 0.0351 0.0191 -0.1706 0.1263 0.2509 -0.0873 1 Panel B AB-to-TP CP-to-TP ln(ta) ROA M/B D/E TMR AB-to-TP 1 CP-to-TP -0.5885 1 ln(ta) -0.0167 0.1723 1 ROA 0.1830 -0.0651 -0.0642 1 M/B -0.0312 0.0961 -0.0919 0.0944 1 D/E -0.1240 0.0353 0.3141 -0.1560 0.0120 1 TMR 0.0323 0.0240 -0.1706 0.1263 0.2509 -0.0873 1 Panel C AB-to-TA CP-to-TA ROA M/B D/E TMR AB-to-TA 1 CP-to-TA 0.1991 1 ROA 0.1879 -0.0778 1 M/B 0.0278 0.2189 0.0944 1 D/E -0.1863 -0.0948 -0.1600 0.0120 1 TMR 0.1495 0.1984 0.1263 0.2509 -0.0873 1 Notes: 1. data include widely-held firms. 2. ln(AB) denotes natural log of annual bonus, ln(CP) denotes natural log of contingent pay, ln(TA) denotes natural log of total assets, ROA denotes return on assets, M/B denotes market to book ratio, TMR denotes total market return ratio, and D/E denotes debt to equity ratio.

86

Table 16.1: The impact of ownership structure on the natural log of annual bonus (OLS) In model9, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 10, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 11, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 12, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 9 T-stat 10 T-stat 11 T-stat 12 T-stat Α -1.5454 -0.5130 -3.3902 -1.0523 -3.4299 -1.0652 -19.6036 -4.6953*** Ln(TA) 0.5168 3.7199*** 0.6051 4.0535*** 0.6106 4.1064*** 1.4112 7.3240*** ROA 0.1566 5.2311*** 0.1521 4.8853*** 0.1486 4.8407*** 0.0205 0.3876 TMR 0.4111 1.0525 0.5052 1.2611 0.4878 1.2181 0.6854 1.3182 D/E -0.0032 -1.5048 -0.0031 -1.4458 -0.0031 -1.4582 -0.0112 -3.9489*** Dyr04 0.8465 1.2155 0.8508 1.2178 0.9009 1.2823 0.9274 1.3685 Dyr05 1.2694 1.9345* 1.2448 1.8939* 1.3654 2.0735** 1.4945 2.3406** Dyr06 1.4263 2.1523** 1.4076 2.0981** 1.4497 2.1546** 1.5865 2.4680** Dyr07 1.1932 1.7440* 1.1692 1.7017* 1.1513 1.6713* 1.3898 2.0571** Dfin -0.7448 -1.0436 -0.7786 -1.0908 -1.6228 -2.1277** Degy -0.3771 -0.6636 -0.4177 -0.7326 -0.6913 -1.1919 Dmat 0.5131 1.0361 0.5134 1.0366 -0.5893 -1.1202 D1yc -1.2368 -1.3439 -1.2003 -1.3737 Dcon 31.1011 4.9057*** Dcon*ln(TA) -1.4748 -5.0489*** Dcon*ROA 0.2142 3.3102*** Dcon*TMR -0.5022 -0.6543 Dcon*D/E 0.0082 2.2521** Adjusted R-square 0.086 0.087 0.089 0.148 F-statistic 7.597 5.821 5.545 6.717 Observations 560 560 560 560

87

Table 16.2: Natural log of annual bonus in widely-held, institution-controlled, and family-controlled firms (OLS) In model 13, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 13 T-stat 14 T-stat α -20.8925 -4.9475*** 13.5344 2.2707** Ln(TA) 1.4690 7.5444*** -0.1466 -0.5300 ROA 0.0187 0.3553 0.2037 5.1417*** TMR 0.6552 1.2639 -0.0527 -0.0870 D/E -0.0111 -3.9212*** -0.0078 -1.6359 Dyr04 0.9991 1.4815 0.9991 1.4815 Dyr05 1.4806 2.3224** 1.4806 2.3224** Dyr06 1.5273 2.3941** 1.5273 2.3941** Dyr07 1.3596 2.0257** 1.3596 2.0257** Dfin -1.9181 -2.4531** -1.9181 -2.4531** Degy -0.3204 -0.5484 -0.3204 -0.5484 Dmat -0.3792 -0.6973 -0.3792 -0.6973 D1yc -1.1821 -1.3776 -1.1821 -1.3776 DF 24.4537 2.5635** -9.9732 -0.9122 DF*ln(TA) -1.2197 -2.8614*** 0.3960 0.8000 DF*ROA 0.4749 4.4327*** 0.2899 2.8530*** DF*M/B 2.1278 0.9085 2.8356 1.2057 DF*D/E 0.0094 2.5068** 0.0061 1.1347 DI 34.4269 4.4862*** DI*ln(TA) -1.6156 -4.5090*** DI*ROA 0.1850 2.8174*** DI*M/B -0.7079 -0.9026 DI*D/E 0.0033 0.6055 Adjusted R-square 0.159 0.159 F-statistic 5.818 5.818 Observations 560 560

88

Table 17.1: The impact of ownership structure on the natural log of contingent pay (OLS) In model9, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. In model 10, we add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). In model 11, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. In model 12, we combine family and institution as the concentrated group (Dcon) to compare to widely-held group. We use the widely-held group as the base group. Dcon is the dummy variable that takes the value of 1 if a firm is concentrated group and 0 otherwise. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 9 T-stat 10 T-stat 11 T-stat 12 T-stat α -3.9256 -1.4493 -13.292 -3.908*** -13.2629 -3.8871*** -27.4721 -6.1060*** Ln(TA) 0.6777 5.3508*** 1.106 6.940*** 1.1022 6.8942*** 1.7839 8.5514*** ROA -0.0321 -1.0074 -0.066 -1.989** -0.0637 -1.9071* -0.0894 -1.9506* TMR 0.6987 1.6289 0.738 1.695* 0.7505 1.7213* 0.6358 1.2311 D/E -0.0004 -0.2007 -0.001 -0.350 -0.0008 -0.3449 -0.0061 -1.9478* Dyr04 0.9582 1.2148 0.885 1.145 0.8475 1.0964 0.9637 1.2621 Dyr05 1.4713 1.9590* 1.339 1.838* 1.2494 1.6954* 1.3972 1.9218* Dyr06 0.8659 1.0956 0.686 0.892 0.6543 0.8508 0.7993 1.0399 Dyr07 1.1827 1.4853 0.943 1.196 0.9562 1.2099 1.2284 1.5348 Dfin -2.336 -2.752*** -2.3106 -2.7028*** -3.0148 -3.3609*** Degy 1.606 2.502** 1.6366 2.5447** 1.5676 2.4021** Dmat 2.254 3.995*** 2.2541 3.9838*** 1.4669 2.5027** D1yc 0.9196 1.0860 0.8981 1.0910 Dcon 27.0490 4.5100*** Dcon*ln(TA) -1.2608 -4.5493*** Dcon*ROA 0.0353 0.5407 Dcon*TMR 0.4093 0.4802 Dcon*D/E 0.0049 1.1771 Adjusted R-square 0.041 0.087 0.087 0.109 F-statistic 3.958 5.837 5.437 5.038 Observations 560 560 560 560

89

Table 17.2: Natural log of contingent pay in widely-held, institution-controlled, and family-controlled firms (OLS) In model 13, use total assets (ln (TA)), return on assets (ROA), total market return (TMR), debt to equity ratio (d/e), and year effects as our independent variables. We use year 2003 as the base year and add the four dummy variables Dyr04, Dyr05, Dyr06, and Dyr07 to represent respectively 2004, 2005, 2006, and 2007. We also add industry effects: dummy finance industry (DFin), dummy energy industry (DEgy), and dummy material industry (DMat). Moreover, we add a dummy variable (D1YC) that takes the value of 1 if there is a CEO change in the year and 0 otherwise. The sample is divided between widely-held, institution-controlled, and family controlled with the widely-held group serving as the base group. DF is the dummy variable that takes the value of 1 if a firm is family-controlled and 0 otherwise, while DI is a dummy variable that takes the value of 1 if a firm is institution-controlled and 0 otherwise. Also, we add variables which are DF multiples each independent variables, and DI multiples each independent variables. In model 6, we use the institution-controlled group (DI) as the base group and keep all other variables of the model 5 unchanged. Note: ***, **, and * denote respectively significance at the 1%, 5%, and 10% levels. Independent Variable Model 13 T-stat 14 T-stat α -27.5014 -6.0868*** 3.7031 0.6546 Ln(TA) 1.7852 8.5409*** 0.3380 1.2566 ROA -0.0901 -1.9578* -0.0270 -0.5047 TMR 0.6211 1.1990 0.8498 1.1897 D/E -0.0062 -1.9695** -0.0032 -0.7399 Dyr04 0.9442 1.2255 0.9442 1.2255 Dyr05 1.4785 2.0200** 1.4785 2.0200** Dyr06 0.8559 1.1061 0.8559 1.1061 Dyr07 1.2802 1.6018 1.2802 1.6018 Dfin -2.9918 -3.3859*** -2.9918 -3.3859*** Degy 1.6006 2.3313** 1.6006 2.3313** Dmat 1.3042 2.1974** 1.3042 2.1974** D1yc 0.7819 0.9557 0.7819 0.9557 DF 19.1567 2.0714** -12.0478 -1.1966 DF*ln(TA) -0.9084 -2.1897** 0.5388 1.1725 DF*ROA -0.1157 -0.7787 -0.1788 -1.1791 DF*M/B 0.5961 0.1725 0.3673 0.1060 DF*D/E 0.0053 1.0954 0.0022 0.3961 DI 31.2045 4.4064*** DI*ln(TA) -1.4473 -4.3301*** DI*ROA 0.0631 0.8968 DI*M/B 0.2287 0.2635 DI*D/E 0.0031 0.5622 Adjusted R-square 0.108 0.108 F-statistic 4.079 4.079 Observations 560 560

90

Appendix II: Figures Figure 1.1: Salary in monetary terms paid by permanent CEO firms: comparison between

widely-held, institution-controlled, and family controlled firms (Adjusted for inflation)

Figure1.2: Bonus in monetary terms paid by permanent CEO firms: comparison between widely-

held, institution-controlled, and family controlled firms

0

100000

200000

300000

400000

500000

600000

700000

800000

900000

1000000

2003 2004 2005 2006 2007

Family

Institution

Widely Held

0

500000

1000000

1500000

2000000

2500000

3000000

2003 2004 2005 2006 2007

Family

Institution

Widely Held

91

Figure1.3: Contingent compensation in monetary terms paid by permanent CEO firms:

comparison between widely-held, institution-controlled, and family controlled firms

Figure1.4: Total compensation in monetary terms paid by permanent CEO firms: comparison

between widely-held, institution-controlled, and family controlled firms

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000

4500000

5000000

2003 2004 2005 2006 2007

Family

Institution

Widely Held

0

1000000

2000000

3000000

4000000

5000000

6000000

7000000

8000000

9000000

2003 2004 2005 2006 2007

Family

Institution

Widely Held

92

Figure 2.1: Salary as a percentage of total compensation paid by permanent CEO firms:


Figure 2.2: Annual Bonus as a percentage of total compensation paid by permanent CEO firms:


0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

2003 2004 2005 2006 2007

Family Bonus

Institution Bonus

Widely-held Bonus

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

2003 2004 2005 2006 2007

Family Salary

Institution Salary

Widely-Held Salary

93

Figure 2.3: Contingent Compensation as a percentage of total compensation paid by permanent

CEO firms: comparison between widely-held, institution-controlled, and family controlled firms

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

2003 2004 2005 2006 2007

Family Contingent Pay

Institution Contingent Pay

Widely-held Contingent Pay

94

Figure 3.1: Changes in compensation following CEO turnovers in family-controlled firms

Figure 3.2: Changes in compensation following CEO turnovers in institution-controlled firms

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

salary bonus contingent other total

New

Old

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000

4500000


New

Old

95

Figure3.3: Changes in compensation following CEO turnovers in widely-held firms

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000


New

Old

96

Figure 4.1: Changes in compensation following CEO retirements in family-controlled firms

Figure 4.2: Changes in compensation following CEO retirements in institution-controlled firms

Notes: retirement firms are firms with paying a large amount of retirement fee to CEOs.

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000


New

Old

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

4000000

4500000


New

Old

97

Figure 5.1: Total compensation as a function of total assets (all data)

Notes: The horizontal axis is total assets and the vertical axis is total compensation. The unit of the X-axis is $1 billion and the unit of Y-axis is 1 million.

0

10

20

30

40

50

60

70

80

$0 $100 $200 $300 $400 $500 $600 $700

Total Compensation as Function of Total Assets


98

Figure 5.2: Total compensation as a function of total assets (asset sizes of $20.48 million – 9

billion)


0

2

4

6

8

10

12

14

16

18

20

$0 $1 $2 $3 $4 $5 $6 $7 $8 $9 $10



99

Figure 5.3: Total compensation as a function of total assets (asset sizes of $9.1 billion – $55 billion)


0

10

20

30

40

50

60

70

80

$0 $10 $20 $30 $40 $50 $60



100

Figure 5.4: Total compensation as a function of total assets (asset sizes larger than $55 billion)


0

5

10

15

20

25

30

$0 $100 $200 $300 $400 $500 $600 $700



101

Figure 6.1: Salary and Bonus are as a function of total assets (all data included)

Notes: The horizontal axis is total assets and the vertical axis is salary and bonus. The unit of the X-axis is $1billion and the unit of Y-axis is 1 million.

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

0 100 200 300 400 500 600 700

Salary and Bonus as Function of Asset Size


102

Figure 6.2: Salary and Bonus are as a function of total assets (asset sizes of $20.48 million - $55

billion)

Notes: The horizontal axis is total assets and the vertical axis is salary and bonus. The unit of the X-axis is $1billion and the unit of Y-axis is 1 million.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

0 10 20 30 40 50 60



Ownership Structure and Executive Compensation in Canadian ... · Ownership Structure and Executive...

Documents

Transcript of Ownership Structure and Executive Compensation in Canadian ... · Ownership Structure and Executive...